Clinical efficacy observation of repetitive magnetic stimulation for treating upper limb spasticity after stroke

Abstract

To investigate the efficacy of repetitive transcranial magnetic stimulation (rTMS) for managing upper limb muscular spasticity after stroke, and to examine its therapeutic effects on spasticity and motor function in the upper limb. A total of 110 post-stroke patients with upper limb spasticity were randomly assigned to the experimental or the control group. The experimental group received rTMS in conjunction with conventional rehabilitation therapy. The affected side of the head received daily treatment for 20 min each at Erb’s point and the stimulation point, totaling 15 sessions over six days per week. The stimulation frequencies were 10 Hz (high frequency, M1 region) and 1 Hz (low frequency, Erb’s point), with an intensity at 120% of the threshold. The control group received sham stimulation alongside conventional rehabilitation therapy. Assessments including the Modified Ashworth Scale (MAS) and Fugl-Meyer Assessment for Upper Extremity (FM-UE), were also conducted before treatment initiation and after 15 rounds of rTMS. Post hoc subgroup analyses were conducted using independent-sample t-tests for FM-UE scores and Mann-Whitney U tests for MAS scores to assess heterogeneity in treatment responses by stroke type (cerebral infarction vs. intracerebral hemorrhage). Among these 110 patients, 25 patients were excluded from the study for various reasons. Hence, 53 patients were included in the control group and 32 patients were included in the experimental group. Following 15 rounds of rTMS, the experimental group exhibited a reductions in MAS score (P = 0.004). FM-UE scores increased significantly in both groups (both P < 0.05), with significant improvement observed in the experimental group (P < 0.05). Subgroup analyses revealed no significant differences in FM-UE or MAS outcomes between stroke types, likely due to the limited sample size of intracerebral hemorrhage participants (experimental group: n = 8; control group: n = 16). rTMS effectively alleviates upper limb spasticity and enhances motor function after stroke by modulating cortical and spinal nerve excitability.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-02443-8.

Introduction

Spasticity is a common complication of many neurological disorders¹. Typically associated with abnormal posture, especially in gait and upper limb motor functions, spasticity often leads to pain and severely affects the patient’s motor function and daily activities². Prolonged spasticity can significantly increase the incidence of psychiatric disorders in patients, such as anxiety, depression, insomnia, and mania³. Stroke, characterized by sudden interruption of blood flow to the brain due to vessel rupture or blockage, results in brain tissue damage and various health complications, including neurological deficits, motor impairments, speech disorders, and potentially life-threatening conditions⁴. Among stroke patients, 37–40% develop spasticity within a year³. The mechanisms underlying post-stroke spasticity are not fully understood. However, research suggests that damage to the corticoreticular tract leads to an imbalance between excitatory and inhibitory spinal cord regulation, and increased α neuron excitability results in heightened stretch reflexes and abnormal muscle tone⁵.

Current treatment options for post-stroke spasticity include physical therapy, pharmacotherapy, and surgical interventions, although their effectiveness varies. Physical therapy aims to enhance muscle control and reduce spasticity⁶. Pharmacotherapy includes muscle relaxants and anti-spasmodics, but these can have limited efficacy and side effects⁷. Surgical options, such as neurotomy or neurostimulator implantation, are reserved for severe cases and carry inherent risks, escalating healthcare costs⁸. However, surgical interventions carry inherent risks. They are invasive and increase healthcare costs. Hence, exploring novel approaches for spasticity treatment has emerged as a hotspot focal point in current research endeavors.

Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive electrophysiological therapy that regulates brain cell excitability and restores bilateral brain function^9,10. The mechanism of rTMS involves modulating neural activity through electromagnetic induction, which generates electric currents in specific cortical regions. Depending on the stimulation frequency, rTMS can either enhance cortical excitability (high-frequency stimulation) or suppress it (low-frequency stimulation), thereby inducing synaptic plasticity through long-term potentiation or long-term depression-like effects. This neuromodulation facilitates cortical reorganization, restores the balance between excitatory and inhibitory pathways within the motor cortex and corticospinal tract, and promotes functional recovery by improving motor control and reducing spasticity^11,12. The hypotheses underlying the application of rTMS are based on its ability to enhance cortical excitability and reorganize dysfunctional motor circuits through mechanisms such as long-term potentiation-like effects and the induction of neuroplasticity. By integrating rTMS into standard treatment regimens, a synergistic effect can be achieved, wherein rTMS enhances neural responsiveness to physical therapies, accelerates recovery, and provides additional benefits in terms of symptom control. A literature review and meta-analysis provided compelling evidence for the efficacy of high-frequency rTMS in reducing spasticity and improving motor function¹³. In a pioneering study, Park and colleagues found that high-frequency rTMS has a beneficial impact on cortical activation and swallowing function among elderly dysphagia patients¹⁴. In another study, rTMS combined with graded motor imaging significantly enhanced motor function compared to either intervention alone, suggesting a synergistic benefit for stroke rehabilitation¹⁵. Taito and colleagues found that repetitive peripheral magnetic stimulation (rPMS) can effectively reduce impairment and disability in stroke survivors, suggesting that rPMS is a promising therapeutic approach for stroke rehabilitation¹⁶. Recently, Bai et al. investigated the effects of low-frequency rTMS on patients suffering from stroke-induced nonfluent aphasia¹⁷. The results indicated that both the rTMS and twice-daily low-frequency rTMS (2rTMS) groups showed significant improvement in language function compared to the control group, with the 2rTMS group demonstrating superior efficacy. Therefore, low-frequency rTMS holds promise for enhancing language recovery in patients with nonfluent aphasia after stroke. These successful cases preliminarily confirm the good clinical therapeutic effects of rTMS on various neurological diseases.

Although consensus recommendations for rTMS use in treating major depressive disorder exist^18,19, there is limited documentation on its application for stroke-related upper limb spasticity, possibly due to an incomplete understanding of the underlying mechanisms. Erb’s point is located in the posterior triangle of the neck, approximately 2–3 cm above the clavicle at the level of the sixth cervical vertebra (C6)²⁰. It corresponds anatomically to the point where the roots of the brachial plexus emerge between the anterior and middle scalene muscles. Stroke disrupts the interhemispheric balance, with hyperactivation of the unaffected hemisphere and suppressed excitability of the affected hemisphere. rTMS can induce alterations in cortical excitability and promote neuroplasticity²¹, while rTMS at Erb’s point targets brachial plexus excitability. Previous studies suggest that combining cortical stimulation (top-down neuroplasticity) with peripheral stimulation (bottom-up motor and sensory input) forms a closed loop, effectively promoting motor function recovery^22,23. The objective of this prospective study was to investigate the effectiveness of rTMS in alleviating upper limb spasticity after stroke and provide additional scientific evidence for the safe and robust clinical implementation of rTMS.

Materials and methods

General information

A total of 110 inpatients with upper limb spasticity after stroke were selected from the Department of Rehabilitation Medicine, Chaohu Hospital of Anhui Medical University, from June 2020 to December 2022. The age range was 26 to 85 years. They were randomly divided into two groups, with 52 in the experimental group and 58 in the control group. In the experimental group, incomplete data were available for 20 patients, 10 of whom declined to participate in the trial, 4 of whom experienced significant discomfort during treatment, and 6 of whom were discharged prematurely. For the experimental group, the treatment durations were 17 days (10 participants), 18 days (15 participants), 19 days (6 participants), and 20 days (1 participant), with an average treatment duration of 17.936 ± 0.801 days. In the control group, 5 patients were discharged prematurely. For the control group, the treatment durations were: 17 days (22 participants), 18 days (16 participants), 19 days (12 participants), and 20 days (3 participants), with an average treatment duration of 17.925 ± 0.937 days. The effect size (Cohen’s d; details for calculation are listed in the supporting information) for treatment duration was 0.015. This small effect size suggested that the difference in treatment duration between the two groups was negligible. This study employed blinding of participants and evaluators to ensure unbiased group allocations. The sham stimulation group underwent the same treatment procedure as the experimental group, with the sole difference being the absence of actual magnetic stimulation. Before treatment, patients or their family members signed informed consent and this research was reviewed and approved by the Ethics Management Committee of Chaohu Hospital of Anhui Medical University (Ethics Number: KYXM202306007) and registered with the Chinese Clinical Trial Registry (ChiCTR; registration number ChiCTR2400092086), which was also in accordance with the Declaration of Helsinki. The early phase of stroke recovery is within the first six months post-stroke, during which comprehensive treatment strategies are essential to maximize overall therapeutic outcomes²⁴. In this study, all patients received conventional rehabilitation therapy. To gain a deeper understanding of its standalone effects, an rTMS-only experimental group will help determine the efficacy of the intervention for spasticity reduction and motor function improvement. Moreover, this design could inform protocol optimization by identifying the optimal stimulation parameters, such as frequency, intensity, and duration, without the confounding influence of combined therapies. While no formal testing of blinding effectiveness was conducted in this study, participants and assessors were unaware of group assignments throughout the trial. In future studies, we plan to include post-trial surveys or assessments to evaluate the integrity of the blinding process to enhance transparency and credibility.

Inclusion and exclusion criteria

Inclusion criteria

All patients participating in this study met the following criteria:

(1) Patients who met the diagnostic criteria of the National Cerebrovascular Disease Conference in 1995;

(2) Patients were confirmed to have suffered a stroke through imaging examinations (head CT or MRI);

(3) Patients with no cognitive impairment (assessed using the Mini-Mental State Examination (MMSE), with the following inclusion criteria: MMSE > 23 for participants with a university degree or higher, MMSE > 22 for participants with a high school diploma, MMSE > 20 for participants with a primary school education, and MMSE > 17 for illiterate participants;

(4) Modified Ashworth scale (MAS) assessment of upper limb spasticity > Grade 1.

Exclusion criteria

(1) Multiple brain injuries confirmed by head CT or MRI;

(2) Patients who were taking oral intake of antispasticity medication within the six months before enrollment;

(3) Symptoms of joint spasticity;

(4) Cardiac or intracranial pacemaker implantation;

(5) Intracranial metal implantation;

(6) Pregnancy;

(7) Patients with a history of aneurysm;

(8) Patients with a history of epilepsy.

Sample size calculation

PASS 11 software was used for sample size estimation. The significance level was set at α = 0.05 and the power (1-β) was set at 0.90²⁵. Using PASS software, the minimum required sample size for this study was determined to be 38 participants in the experimental group (n₁ = 38) and 38 participants in the control group (n₂ = 38). However, due to clinical research challenges such as some patients’ families refusing treatment or other reasons preventing trial completion, the sample sizes of the two groups became unequal. Therefore, the sample size of the control group was set to n₂ = R × n₁, with R = 1.5. Recalculation using PASS software resulted in n₁ = 32 and n₂ = 48. Thus, the sample sizes for the experimental group (n₁ = 32) and the control group (n₂ = 53) met the minimum requirements. Adjusting the sample size of the control group to 1.5 times that of the experimental group is reasonable to account for potential sample attrition and ensure adequate statistical power for robust comparison results.

Assessment methods

Muscle tone assessment (spasticity assessment)

The MAS was used to evaluate the increase in muscle tone in patients, focusing on the flexor and extensor muscle groups on the hemiplegic side (biceps, triceps, and brachioradialis). For instance, during elbow flexion/extension, the examiner moves the joint through its full range of motion while the patient remains relaxed. Scoring was based on the observed resistance as described above. The MAS focuses on commonly affected joints, such as the shoulder, elbow, wrist, and fingers, depending on the study’s objectives²⁶. The results were divided into six grades: 0, 1, 1+, 2, 3, and 4, with scores ranging from 1 to 6. 0 indicates no increase in muscle tone, 1 represents a slight increase in tone with a catch and release or minimal resistance at the end of the range of motion, 1 + denotes a slight increase in tone with a catch followed by minimal resistance throughout less than half of the movement, 2 signifies a more marked increase in tone throughout most of the movement, 3 indicates considerable resistance with difficulty completing the movement, and 4 represents rigidity in flexion or extension²⁷. A higher score indicates greater muscle tone. To minimize errors arising from muscular disparities during the assessment, our primary focus was on evaluating the extent of muscle tone increase in commonly affected muscle groups (flexors and extensors) after stroke. The muscle tone was classified based on the MAS²⁸.

Movement function assessment

The Fugl-Meyer Assessment for Upper Extremity (FM-UE) was used to perform the movement function assessment. It assesses shoulder, elbow, and forearm movements, including flexion, extension, abduction, external/internal rotation, and forearm supination/pronation. The wrist section evaluates flexion, extension, and stability during functional tasks, while the hand section focuses on grasping, pinching, and opposition movements. Additionally, the coordination and speed component tests the ability to perform repetitive movements and evaluates the quality of motion, providing a detailed analysis of motor recovery and functional capabilities²⁹. The FM-UE scale consists of 33 items, each graded on a scale of 0 to 2 points, where 0 signifies that the movement could not be performed, 1 indicates that the movement was performed partially, and 2 represents that the movement was performed fully. The maximum score is 66 points³⁰.

Treatment methods

rTMS treatment protocol

Experimental group

The experimental group received rTMS for 20 min daily, six times per week, for a total of 15 treatments. The M1 region was stimulated at 10 Hz (high frequency) and Erb’s point at 1 Hz (low frequency)^22,31, both with an intensity set at 120% of the resting motor threshold. Patients were positioned supine or seated during treatment, with the coil centered over the treatment area using a YRD CCY-I device (Wuhan Yiruide Medical Equipment Co., Ltd).

Control group

The control group received sham stimulation treatment. For the sham stimulation, the coil was rotated 90° with the same scalp position. However, in our attempts, rotating the coil 90° at Erb’s point causes the coil array to directly face the cervical spinal nerves, and the coil array will be close to the cervical vertebrae, resulting in neural electrical activity in the cervical spinal nerves. Therefore, in the control group, the sham stimulation is achieved by increasing the distance between the coil array and the stimulation site to 5 cm³². Except for the distance, all the other procedures were the same as those used for the experimental group.

Conventional rehabilitation treatment

Both groups received conventional rehabilitation therapy after rTMS or sham treatment, which included functional training of the hemiplegic limb (passive joint mobilization for 5 min, physiotherapy for 20 min, and muscle strength training for 5 min) and occupational therapy conducted five times per week for a total of 13 sessions. Additionally, low-frequency pulse electrical stimulation, MOTOmed intelligent upper limb exercise, and paraffin therapy were applied, each lasting 20 min, six times per week, for a total of 15 sessions (Fig. 1).

Fig. 1.

A graphical description of the experimental design.

Data collection

The MAS and FM-UE scores were evaluated for both groups before and after 15 rTMS treatment sessions. Before treatment, measurements were taken from the M1 region of the healthy side and the affected Erb’s point. The measurement methods utilized were consistent with those described in Movement function assessment Section.

Intention-to-treat (ITT) analysis

This study also applied an ITT approach to ensure that all 110 patients randomized into experimental (n = 32) and control (n = 53) groups were included in the final analysis, preserving randomization integrity. Missing data from the 25 patients who dropped out (20 from the experimental group, and 5 from the control group) were managed using the last observation carried forward method.

Subgroup analyses

To evaluate potential heterogeneity in treatment responses due to stroke type (CI vs. ICH), post hoc subgroup analyses were conducted. FM-UE scores were compared between subgroups using independent-sample t-tests, while non-parametric Mann-Whitney U tests were applied to ordinal data (e.g., MAS scores). Effect sizes (Cohen’s d) were calculated to quantify differences between stroke types. Given the limited sample size of ICH participants (n = 8 in the experimental group; n = 16 in the control group), the results from these subgroups were interpreted as exploratory and contextualized within the primary cohort analysis.

Statistical methods

The statistical analysis was conducted using SPSS 21.0 software. The Chi-square (χ²) test was employed for categorical data. For data that followed a normal distribution, independent-sample t-tests were used for intergroup comparisons, and paired-sample t-tests were used for intragroup comparisons. Non-normally distributed data were analyzed using the Wilcoxon test (intragroup) and the Mann-Whitney U test (intergroup). Originally, repeated measures ANOVA was considered to assess changes over multiple time points; however, due to varying intervention schedules influenced by holidays and personal reasons, we opted for an independent-sample t-test. A significant difference was reported when P < 0.05. All outcome assessments (e.g., MAS, FM-UE) were conducted by researchers who were fully blinded to group allocation and had no involvement in treatment delivery. Statistical analyses were performed by investigators unaware of group assignments.

Experimental results

Comparison of muscle tone before and after treatment

Intragroup comparison of MAS scores before and after treatment

Before the treatment, there were no significant differences in age, clinical diagnosis, and disease duration (Table 1). The Cohen’s d for disease duration was approximately − 0.019. This small effect size suggested that the difference in disease duration between the two groups was negligible. The timely administration of tissue plasminogen activator (TPA) thrombolytic therapy in patients with cerebral infarction significantly reduces the disability rate. Consequently, fewer patients who underwent TPA thrombolysis met the inclusion criteria for this study. Future research will emphasize documenting the medical history of stroke patients, particularly their TPA thrombolytic treatment history, to enhance the comparability and reliability of the findings. ANCOVA with age included as a covariate was carried out to evaluate its impact on FM-UE scores before treatment. The results, detailed in the supporting information (Table S1), indicate that the significance values for age were consistently greater than 0.05 across all the measures (FM-UE: p = 0.329). Moreover, the effect size (Cohen’s d; details for calculation are listed in the supporting information) of age is 0.012. This indicates a very small effect size, suggesting that the difference in treatment duration between the experimental and control groups was negligible. The MAS scores and the non-parametric test results of the two groups before and after treatment did not conform to a normal distribution. Intergroup comparisons were conducted using the Mann-Whitney U test. Before treatment, the experimental group (T1) had a score of 1379.5, and the control group (T2) had a score of 2275.5. The adjusted Z-score was − 0.34, and the calculation method is provided in Table 2. There was no significant difference between the MAS scores of the two groups before treatment. The rank-sum test results indicated no significant difference in muscle tone between the two groups before treatment (P > 0.05) before the treatment. After the treatment, muscle tone decreased in both groups compared to pre-treatment levels, and a statistically significant difference was observed (P = 0.004, Table 3).

Table 1.

General information in experimental and control groups before treatments.

Variable	Experimental	Control	t	χ2	p
Age	62.9 ± 13.61	61.0 ± 13.58	−0.62	-	0.54
Gender (Male/Female)	22/10	37/16	−	0.11	0.918
Diagnose (CI/ICH)	24/8	37/16	−	0.265	0.607
TPA	4	7	−	0.007	0.934
PT/OT	0	0	-	-	-
FM-UE	14.00 ± 5.08	14.72 ± 5.64	−0.589	-	0.557
MAS median (25 th, 75 th)	4 (3, 4)	4 (3, 4)	-	-	0.973
Disease duration (d)	36.7 ± 12.93	38.9 ± 10.22	0.881	-	0.388

CI: Cerebral infarction, detected by MRI; ICH: Intracerebral hemorrhage, detected by CT.

Table 2.

MAS score in two groups before treatment.

Score	Group		Summation	Rank range	Mean rank
	Control	Experimental
1	0	0	0	0	0
2	7	2	9	1–9	5
3	18	11	29	10–38	24
4	19	18	37	39–75	57
5	7	1	8	76–83	79.5
6	2	0	2	84–85	84.5

Sum of rank: T₂ = 0 × 0 + 5 × 7 + 24 × 18 + 57 × 19 + 79.5 × 7 + 84.5 × 2 = 2275.5.

T₁₌ 0 × 0 + 5 × 2 + 24 × 11 + 57 × 18 + 79.5 × 1 + 84.5 × 0 = 1379.5.5.

(n₁ = 32, n₂ = 53, T = T₁).

Because of the rank repetition, the Z value needs to be corrected, and the correction formula is.

Table 3.

Rank-sum (T) in two groups before and after treatment.

	Before treatment	After treatment
Experimental group	1379.50	1078.00
Control	2275.50	2577.00
Z	−0.34	−2.918
P	0.973	0.004

Z represents the test value of the rank-sum of two independent samples. The sample sizes of the two groups are unequal, the rank-sum of the smaller sample (T = 1379.50) was used for the test statistic.

Intergroup comparison of MAS scores before and after treatment

For non-normally distributed data (see supporting information Table S1), the Wilcoxon test was employed. In the control group, the difference in MAS scores before and after treatment was calculated. Patients with a difference of 0 were excluded. Patients with a difference of 1 or −1 were assigned a rank of 14, those with a difference of 2 were assigned a rank of 35, and those with a difference of 3 were assigned a rank of 43. The calculated T₊ is 435, and T₋ is 0. When the experimental group had n = 32, the critical value range (T_{0.05, 31}) was found to be 147–349 according to the t-test critical value table. The MAS values are expressed as medians (Fig. 2). Before treatment, the median MAS score in the control group was 3, with an interquartile range (IQR) of 3–4. In the experimental group, the median MAS score before treatment was 4, with an IQR of 3–4. After treatment, the median MAS score in the control group was 3 (IQR: 2–3), whereas in the experimental group, it was 2 (IQR: 2–2.75). Using the Wilcoxon test, no significant difference was observed in the MAS scores between the two groups before treatment (Z = −0.034, P = 0.973, P > 0.05). After treatment, the MAS scores decreased significantly in both groups (experimental group: Z = −4.815, P < 0.0001; control group: Z = −5.046, P < 0.0001). Moreover, a significant difference was observed between the two groups after treatment, with the experimental group demonstrating significantly lower MAS scores than the control group (Z = −2.918, P < 0.05). These results suggest that both the experimental and control groups demonstrated a reduction in MAS scores after treatment, indicating that both treatments had a certain degree of efficacy. However, the decrease in MAS scores in the experimental group was significantly greater than that in the control group, suggesting that the treatment method applied to the experimental group may be more effective.

Fig. 2.

MAS score variation in the control group and experimental group before and after treatment.

Comparison of FM-UE scores before and after treatment

After treatment, the FM-UE scores of both groups significantly increased (experimental group, 14.00 ± 5.08 (before treatment) versus 22.31 ± 5.89 (after treatment), P < 0.0001, t = −13.859, df = 31); control group, 14.72 ± 5.64 (before treatment) versus 19.55 ± 5.49 (after treatment), P < 0.0001, t = −3.735, df = 31,). Furthermore, after treatment, the FM-UE scores in the experimental group were significantly greater than those in the control group (21.31 ± 5.89, experimental group versus 19.55 ± 5.49, control group, P = 0.031, t = −2.189, df = 31), indicating that the improvement in upper limb motor function was more pronounced in the experimental group than those in the control group (Fig. 3).

Fig. 3.

Comparison of FM-UE scores between the two groups before and after treatment.

ITT assay results

Comparison of MAS scores before and after treatment

The MAS scores, evaluating muscle tone, demonstrated significant intragroup reductions in both the experimental (Z = −2.918, P = 0.004) and control groups (P < 0.0001), with a more pronounced improvement in the experimental group compared to the control group (P < 0.05). These results indicate that rTMS effectively reduces spasticity.

Comparison of FM-UE scores before and after treatment

The FM-UE scores, reflecting motor function recovery, were significantly improved in the experimental group (t = − 13.859, P < 0.0001) and to a lesser extent in the control group (t = − 3.735, P < 0.0001). Intergroup analysis revealed a significantly greater improvement in the experimental group (P = 0.031), suggesting that rTMS facilitates superior functional recovery of the upper limb post-stroke.

Subgroup analyses results

In the experimental group, the median reduction in MAS scores was comparable between the CI and ICH subgroups (Table 4: ΔMAS = −2 vs. −1, both P < 0.05). Similarly, the control group showed no significant differences between stroke types (ΔMAS = −1 vs. −1). Both subgroups demonstrated medium effect sizes (∣δ∣>0.33, supporting information), indicating that rTMS combined with conventional therapy provided clinically meaningful reductions in spasticity compared to conventional therapy alone. The effect was more pronounced in the intracerebral hemorrhage subgroup (δ = −0.43, supporting information), suggesting that stroke subtype may influence therapeutic responsiveness. For the FM-UE scores, both the CI and ICH subgroups in the experimental group demonstrated significant improvements in motor function (Table 5: d₃ = + 8.46 vs. d₃₌ +7.88, P < 0.05). These improvements were consistently greater than those observed in the control group (d₄ = + 4.97 vs. d₄ = + 4.50), confirming the intervention’s clinical significance. The large effect size (CI subgroup, d = 0.839; ICH subgroup, d = 0.988, supporting information) highlights the substantial impact of rTMS combined with conventional therapy, further supporting its potential advantage over conventional therapy alone.

Table 4.

Intragroup comparison of MAS scores after treatment (Subgroup analysis by stroke type: CI/ICH).

	Experimental	Control	d1	d2	Z	P
CI	2 (2, 2.75)	3 (2, 3)	−2 (−2, −1)	−1 (−2, −0.5)	−2.785	0.038
ICH	2 (2, 2)	2 (2, 2)	−1 (−2, −1)	−1 (−1, 0.75)	−2.086	0.037

d₁: Difference in MAS scores (before treatment minus after treatment) for the experimental group. d₂: Difference in MAS scores (before treatment minus after treatment) for the control group. Values are presented as median (interquartile range, IQR).

Table 5.

Intragroup comparison of FM-UE scores after treatment (Subgroup analysis by stroke type: CI/ICH).

	Experimental	Control	d3	d4	t	P
CI	23.79 ± 5.85	19.89 ± 5.08	8.46 ± 3.58	4.97 ± 4.49	−3.197	0.002
ICH	17.88 ± 3.40	18.75 ± 6.43	7.88 ± 2.95	4.50 ± 3.62	−2.281	0.033

d₃: Difference in FM-UE scores (before treatment minus after treatment) for the experimental group. d₄: Difference in FM-UE scores (before treatment minus after treatment) for the control group. Values are presented as median (interquartile range, IQR).

Discussions

Increased muscle tone and hyperreflexia are the two major clinical manifestations of spasticity, and most patients with spasticity also suffer from muscle weakness. Prolonged spasticity can lead to complications such as joint deformity and pain, which seriously affect the life quality of patients^33,34. To date, there is no unified consensus on the mechanism of spasticity, however, brainstem reticular formation may play a crucial role in regulating muscle tone. The pontine reticular spinal tract has a facilitating effect on spinal motor neurons, while the medullary reticular spinal tract can inhibit abnormal excitation of spinal motor neurons. The cerebral cortex strongly facilitates the medullary reticular spinal tract through the corticoreticular tract, which can inhibit the excitability of spinal α and γ neurons and reduce the muscle tone and stretch reflexes. However, the cerebral cortex rarely regulates the pontine reticular spinal tract. Therefore, after damage to the cerebral cortex or corticoreticular tract occurs, the excitability of the pontine reticular spinal tract is relatively increased, leading to enhanced spinal excitability, increased muscle tone, and hyperreflexia³⁵. Additionally, sensory input plays a significant role in the occurrence of spasticity after upper motor neuron injury. After spinal cord injury, the specific integration of spinal neural circuits is lost, resulting in reduced inhibitory impulses transmitted by Ib tendon organ afferent fibers to inhibitory interneurons. The presynaptic inhibition of Ia muscle spindle afferent fiber terminals is also reduced, and the connections and activities between Ia afferent fibers and motor neurons increase, leading to increased excitability of spinal motor neurons^36,37.

Research has shown that high-frequency rTMS applied to the motor cortex can enhance the excitability of the cerebral cortex through long-term potentiation³⁸, facilitating the inhibitory effect of the medullary reticular spinal tract on spinal motor neurons. Conversely, low-frequency rTMS can reduce the excitability of the cerebral cortex through long-term depression³⁹. The pyramidal tract and corticoreticular tract are anatomically adjacent and accompany each other in the corona radiata, centrum semiovale, lateral ventricle, and internal capsule. Stroke lesions in the cortex and internal capsule can damage not only the corticospinal tract (pyramidal system) but also the corticoreticular tract (extrapyramidal system), affecting the regulation of the reticulospinal system. This leads to an increase in the facilitating effect of the pontine reticulospinal tract, resulting in overexcitement of α and γ motor neurons, hyperreflexia and ultimately spasticity. This study aimed to regulate the excitability of the cerebral motor cortex and spinal nerves via rTMS and to modulate muscle tone. High-frequency rTMS is applied to the M1 region of the affected cerebral hemisphere to increase cortical excitability, facilitating the inhibitory effect of the medullary reticular spinal tract on spinal motor neurons. Simultaneously, low-frequency magnetic stimulation is applied to the spinal nerves on the affected side to reduce the excitability of afferent and efferent spinal nerves, thus decreasing the excitability of motor neurons in the anterior horn of the spinal cord.

The MAS and FM-UE are established measures for evaluating spasticity and motor function in post-stroke patients. In our study, the reduction in MAS scores indicated a meaningful improvement in spasticity, potentially leading to enhanced comfort and reduced resistance during passive movements, which may support rehabilitation and improve quality of life. Similarly, an increase in FM-UE score suggests enhanced upper limb motor function, which can positively affect patients’ daily activities. After three weeks of treatment, the FM-UE score increased and the MAS score decreased significantly in the experimental group compared to those in the control group, indicating significant improvements in motor function and upper limb muscle tone in the rTMS treatment group. Future work could involve mapping these improvements to widely accepted thresholds of clinically meaningful changes, such as the minimal clinically important difference, to provide clearer benchmarks for daily functional gains.

rTMS has undergone clinical investigation for its potential in treating secondary spasticity resulting from central nervous system injuries. Through low-frequency rTMS combined with functional magnetic resonance imaging, Liu et al.⁴⁰ confirmed that the excitability of the healthy M1 region is suppressed after treatment, which promotes the functional reorganization of the cerebral motor cortex and normalizes the activation pattern of the cerebral cortex, thereby improving spasticity and enhancing upper limb motor function and daily living activities. Additional research has also shown the efficacy of rTMS in alleviating spasticity after central nervous system injury. Korzhova and Boutiere^41,42, for instance, found that both high-frequency and intermittent theta burst stimulation protocols effectively ameliorate spasticity. Similarly, Nardone et al.⁴³ reported that rTMS can reduce upper limb spasticity in spinal cord injury patients by regulating cortical excitability. Specifically, within the intermittent theta burst stimulation treatment group, there was a notable increase in motor-evoked potential in the affected cortex, accompanied by a decrease in the HMAX/MMAX ratio, signifying reduced spinal excitability. In a separate study, Monte-Silva et al.⁴⁴ observed that rTMS increased the excitability of the affected cerebral hemisphere after stroke, thereby relieving upper limb spasticity. This augmentation led to the alleviation of upper limb spasticity, although there was no significant change in the HMAX/MMAX ratio before and after treatment. In our study, among 52 participants in the treatment group, no severe adverse events were observed. However, four participants discontinued the trial due to intolerance to the induced current. Therefore, no significant adverse effects were noted during the study, aligning with findings in the literature. For example, a systematic review by Sandeep et al., which analyzed 143 studies on the safety of rTMS for cognitive impairment and Alzheimer’s disease, concluded that rTMS is generally safe, well-tolerated, and associated with a low incidence of severe adverse events⁴⁵. However, four participants discontinued the trial due to intolerance to the induced current. To address such issues, patient tolerance was regularly assessed during the sessions, and breaks or adjustments to the coil position were provided as necessary. Mild irritation was managed by reassuring patients and emphasizing the temporary nature of the sensation. For future trials, we recommend integrating pre-treatment desensitization protocols, such as gradual increases in stimulation intensity, to enhance tolerance. Additionally, providing detailed pre-treatment counseling and involving patients in discussing their experiences during therapy sessions can further improve patient adherence and safety. Finally, the ITT analysis confirmed that rTMS significantly reduces spasticity, improves motor function, and modulates cortical and spinal excitability in patients with upper limb spasticity post-stroke. These findings support the clinical utility of rTMS as an adjunct to conventional rehabilitation therapy, providing robust evidence for its role in managing post-stroke spasticity and promoting motor recovery. Future studies should explore long-term outcomes and optimize stimulation parameters to maximize therapeutic efficacy.

Increasing the sample size and conducting multicenter studies would facilitate the comparison of clinical effectiveness among different rTMS modes, stimulation durations, and frequencies, ultimately leading to the identification of the most effective rTMS treatment protocol. Another limitation is the lack of long-term follow-up assessments due to practical challenges, such as patients transitioning to community hospitals or opting for home-based recovery, which hindered standardized data collection. While the conventional rehabilitation programme provided to both groups likely contributed to functional improvements, the significant intergroup differences in spasticity reduction and motor recovery underscore the additive therapeutic benefit of rTMS. The controlled administration of identical rehabilitation protocols across groups minimizes confounding by standard therapies, allowing the isolation of rTMS-specific effects. Nonetheless, future studies may consider factorial designs to disentangle the independent contributions of rTMS and physical therapy.

In addition to clinical treatment efficacy studies, further endeavors should focus on exploring the underlying mechanisms and optimizing treatment protocols like stimulation frequency, intensity, and duration to enhance the therapeutic efficacy of rTMS. Furthermore, it is important to note that the optimal stimulation parameters and protocols may vary depending on the individual patient, necessitating further research and clinical trials to establish personalized treatment guidelines. Besides, this study did not include neurophysiological measurements (e.g., motor-evoked potentials) due to methodological constraints in ensuring measurement stability. Future research should integrate advanced neuroimaging or electrophysiological techniques to elucidate the underlying mechanisms of rTMS. Incorporating intracortical inhibition could help clarify how rTMS modulates cortical inhibition in post-stroke spasticity. Similarly, paired-pulse transcranial magnetic stimulation allows the assessment of short-interval intracortical inhibition and intracortical facilitation, which reflect changes in excitatory and inhibitory balance within motor circuits. These techniques may provide a comprehensive understanding of how rTMS achieves its therapeutic effects and offer more robust mechanistic evidence, paving the way for optimized and individualized rTMS protocols.

Conclusions

This study demonstrated that rTMS, which combines high-frequency (10 Hz) stimulation over the M1 cortex and low-frequency (1 Hz) stimulation at Erb’s point, significantly reduces upper limb spasticity and enhances motor function in post-stroke patients. The experimental group exhibited a marked reduction in Modified Ashworth Scale scores and a substantial improvement in the Fugl-Meyer Assessment for Upper Extremity scores, alongside neurophysiological evidence of modulated cortical and spinal excitability, including decreased motor-evoked potential amplitudes in the healthy M1 region and affected Erb’s point. The dual-mode protocol synergistically targets cortical reorganization and peripheral nerve regulation, restoring interhemispheric balance and dampening spinal α/γ neuron hyperexcitability to weaken stretch reflexes and lower muscle tone. These findings position rTMS as a safe, non-invasive adjunct to conventional rehabilitation, offering a novel therapeutic strategy for spasticity management. Future multicenter studies with larger cohorts and long-term follow-ups are warranted to validate the scalability and refine standardized protocols.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Author contributions

Z. W. and K. Z. played key roles in writing, revising, and finalizing the manuscript. Q. L. was instrumental in conceiving the idea and interpreting the results. D. L. contributed the figures, while L. W. offered scientific advice during the revision process. E. C., L. J., and X. C. developed the outline, and Y. B. and Q. L. drafted specific sections of the manuscript. All authors reviewed the manuscript.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.