Authors: Wenming Lu, Longxiang Yan, Chuangguo Li, Kai Wang, Qijing Wang, Sisi Xu, Benguo Wang
Categories: Review, Repetitive transcranial magnetic stimulation, Parkinson’s disease, Efficacy, Safety, Meta-analysis
Source: Clinical Parkinsonism & Related Disorders
Authors: Wenming Lu, Longxiang Yan, Chuangguo Li, Kai Wang, Qijing Wang, Sisi Xu, Benguo Wang
Parkinson’s disease (PD) is a multifactorial neurodegenerative disease with a high prevalence worldwide, leading to motor and non-motor symptoms. Moreover, PD presents a progressive aggravation alongside time, the middle and late patients in PD requires the use of a variety of anti-Parkinson’s drugs with obvious side effects, which bring serious impact on the quality of life of patients. In recent years, repetitive transcranial magnetic stimulation (rTMS), as a kind of non-invasive neuromodulation therapy, has drawn increasing interest from neurologists, and has been effectively utilized to alleviate both motor and non-motor symptoms of PD. However, the treatment protocols and therapeutic effects of rTMS for PD patients are inconsistent. This meta-analysis aims to systematically evaluate the safety and efficacy of rTMS therapy in patients with PD.
We will perform a comprehensive search in the following electronic PubMed/Medline, Web of Science, EMBASE, and Cochrane, without language restrictions, from their inception to September 2024. This review protocol was formulated according to the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) guidelines. The Cochrane risk of bias tool is utilized to assess the risk of bias. Finally, the effect size was expressed by a standardized mean difference (SMD) with a 95% confidence interval (CI).
A total of 45 randomized controlled trials were included. The results of enrolled studies indicated that both primary and secondary indicators had improved. Subgroup analysis showed that high-frequency rTMS (HF-rTMS) targeting the supplementary motor area (SMD = − 0.56; 95 %CI = [−0.77, −0.36]; p < 0.00001), primary motor cortex (SMD: −1.65; 95 %CI = [−2.35, −0.95]; p < 0.00001), and dorsolateral prefrontal cortex (DLPFC) (SMD: −0.68; 95 %CI = [−1.16, −0.21]; p = 0.005) yielded a significant reduction in motor UPDRS-III scores, compared to the sham group. In addition, HF‐rTMS over left DLPFC or intermittent theta burst stimulation over left DLPFC may benefit cognition. Furtherly, the subgroup analysis of the Beck Depression Inventory scores indicated HF-rTMS over the left DLPFC was a beneficial treatment for depressive symptoms in PD.
The meta-analysis showed that rTMS was effective and safe in the treatment of PD, improving motor function (such as a decrease in UPDRS-III total scores, subscores of UPDRS-III, and FOG-Q scores) in patients with PD and leading to improvement in cognitive function and depression (such as an increase in MocA scores and a decrease BDI scores). Although the results of the subgroup analysis provide a valuable reference for the selection of rTMS for clinical application, further larger multicenter, randomized, placebo-controlled studies with a large number of participants are still required to validate these results.
Parkinson’s disease (PD), a chronic and progressive neurodegenerative disorder, results in heavy global social and health costs [1], [2]. The prevalence of 1 to 2 per 1000 people has been reported [3]. The aging of society is predicted to further increase the population affected by PD in the world which is estimated to reach 12.9 million by 2040 [4], [5]. PD is characterized by various motor dysfunctions, such as bradykinesia, muscle rigidity, gait freezing, resting tremor, and postural instability, which commonly present along with poor balance and coordination, stiffness of the arms, legs, and trunk, and bilateral vocal cord paralysis at the extreme and worsening level [6], [7]. In addition, PD has also a constellation of non-motor symptoms, including constipation, anxiety, dementia and a variety of autonomic dysfunction [8]. These symptoms have been considered valuable for early diagnosis and progression of PD [8], [9].
Currently, clinical treatment strategies of PD mainly consist of medication and deep brain stimulation (DBS). As of right now, levodopa and anticholinergic agents are the most used medication for treating motor symptoms of PD [10]. These have been demonstrated to quickly and effectively alleviate patients' initial symptoms [11]. However, the chronic administration antiparkinsonian drugs cause severe side effects which can aggravate PD symptoms such as involuntary movements and fall [12]. Levodopa, especially, carries several side effects; long-term use of levodopa can result in levodopa-induced dyskinesia and levodopa-resistant; it also does not prevent dopaminergic neuron degeneration and does not affect non-motor symptoms [13], [14]. While, it is reported that DBS induce higher frequency of suicide ideation and suicides [15]. Therefore, it is of great urgency to develop safe and effective treatment agents for PD.
Repetitive transcranial magnetic stimulation (rTMS) is a noninvasive, painless, well-tolerated technique of cortical stimulation based on electromagnetic induction to generate a strong fluctuating magnetic field [16], [17]. It induces intracranial currents and alters cortical activity in the human brain through a stimulation coil placed over the scalp [18], [19]. Cortical excitability is largely dependent upon stimulation parameters of rTMS, such as the frequency and temporal [20]. Recently, rTMS has been adopted into numerous clinical conditions, including depression, migraine, epilepsy, aiding motor recovery in post-stroke patients, and PD [17], [21]. More importantly, rTMS has been investigated as a treatment to improve motor symptoms in PD, including bradykinesia and rigidity, motor complications as well as non-motor symptoms, including depression and speech [22]. In conclusion, rTMS is a promising strategy for PD treatment. Some previous studies have demonstrated the efficacy and safety of rTMS in the treatment of PD. Unfortunately, few of these analyses had specifically focused on the effects of different frequencies and intensities of stimulation in rTMS treatment upon given randomized controlled trials (RCTs). Moreover, clinical studies have been updated in recent years [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57], [58], [59], [60], [61], [62], [63], [64], [65], [66]. Herein, we rigorously screened and extracted data about rTMS against PD in clinical RCTs and aimed to tightly assess the efficacy and safety of rTMS transplantation for PD. Moreover, sensitivity analyses also were conducted to explore sources of heterogeneity between studies. Collectively, our results might provide new insights for elucidating the potential therapeutic role of rTMS in PD.
Our article was conducted by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA, 2020) guidelines, and the protocol was registered on PROSPERO (https://www.crd.york.ac.uk/PROSPERO, registration number is CRD42024614753.
To define studies for inclusion in this meta-analysis, we searched four databases (including PubMed, Embase, Web of Science, and the Cochrane Library) through October 23, 2024, without language restrictions. Databases were searched using the following subject [repetitive transcranial magnetic stimulation] and [Parkinson's disease], and the search was limited to English. Additionally, we searched reference lists of all identified reviews to identify potentially relevant articles. The literature retrieval was conducted independently by one author (Wenming Lu). Potentially eligible articles were also read in full text by two authors (Wenming Lu and Longxiang Yan). After reviewing the full text independently, two authors rechecked the discrepancies. And, if there is a disagreement between two authors, a third author will participate in the discussion until a consensus is reached. The detailed search strategy was presented in the Additional file 2.
Study inclusion was determined independently by two authors (Wenming Lu and Longxiang Yan) based on the screening criteria. A third reviewer was involved in reaching consensus in cases of dispute. The literature search was carried out by combining keywords in the search strategy described above and duplicated articles were excluded. Then, articles that did not meet the established inclusion criteria were excluded by reviewing titles and abstracts. Articles passed the title and abstract screening were finally included in this study.
After identifying the included studies, two authors (Wenming Lu and Pan Cheng) performed data extraction. The following data were extracted from the first author; year of publication; country of publication; type of study design, number of sample sizes for control and experimental groups, primary indicators including Movement Disorder Society–Unified Parkinson’s Disease Rating Scale motor score Part III (UPDRS-III) scores, Beck Depression Inventory (BDI) scores, and Montreal Cognitive Assessment (MoCA) scores, and secondary indicators including freezing of gait questionnaire (FOG-Q) and time up and go (TUG).
The Population, Intervention, Comparison, Outcomes, and Study (PICOS) design model was used to establish the article inclusion Population (P): Patients diagnosed with PD, regardless of country, region, age, sex, and race.Intervention (I): The treatment of the disease was rTMS.Comparison (C): rTMS stimulation groups and the sham stimulation group.Outcomes (O): UPDRS-III scores, subscores of UPDRS-III, MoCA scores, BDI scores, FOG-Q, TUG scores, and adverse events (AE).Study design (S): Randomized controlled trials (RCTs).
The following types of articles were excluded from this
(a)The study did not meet the inclusion criteria; (b) Clinical trials, letters, conference abstracts, animal experiments, meta-analyses, reviews, and case reports; (c) Non-English documents; (d) Single-arm experiment; (e) Inability to access full-text information and extract data for research.
The quality of the included studies was evaluated using Review Manager (version 5.4). All included studies were assessed through seven (1) random sequence generation, (2) allocation concealment, (3) blinding of participants, (4) inadequate outcome data, (5) blinding of outcome assessment, (6) selective reporting, (7) other possible bias. According to the Cochrane Collaboration’s tool for assessing the risk of bias, the studies were classified as “low risk,” “high risk,” or “unclear risk”. Two authors independently assessed the quality of the literature. In cases of doubt, any disagreements were resolved by reaching a consensus or by consulting a third-party author.
This meta-analysis was accomplished via Review Manager version 5.0 software, which was provided by the Cochrane Collaboration. Continuous outcomes were presented as the means and standard deviation (SD). The standardized mean difference (SMD) or weighted mean difference (WMD) with the 95 % confidence interval (CI) for each parameter was calculated to reveal changes after rTMS stimulation, we calculated the mean and SD based on the transformation tools for subsequent analyses. For dichotomous data, relative risk (RR) and odds ratio (OR) were used as the effect size through Mantel-Haenszel (M−H) analysis. We utilized the I^2^ statistic to assess statistical heterogeneity and employed Cochran’s Q test for a qualitative analysis of study heterogeneity. When p > 0.1, I^2^ < 50 %, it is considered that there is no significant heterogeneity between studies, a fixed-effects model was used to pool data. When p ＜ 0.1, I^2^ ＞50 %, the random-effect model is used for analysis. If there was no mention of the mean change, Sensitivity analysis was performed by excluding one study at a time to evaluate the influence of individual studies on pooled results. Subgroup analyses were performed to assess potential sources of heterogeneity. Funnel plots were made to evaluate publication bias and small sample effect. p < 0.05 was considered significant for all outcomes.
In total, 3195 potentially eligible records were obtained from PubMed, Cochrane Library, Web of Science, and Embase databases. After eliminating duplicates, 1825 records remained, and the title and abstracts of these publications were carefully screened. Next, 1694 articles were excluded for the following reasons (with no relevant topics, animal experimental models, not English publications, and reviews). Then, 131 full-text articles were assessed for eligibility criteria. Consequently, 45 trials were included in the systematic review and meta-analysis. The detailed process of the literature search is illustrated in Fig. 1.Fig. 1Flowchart of study selection.
A total of 45 studies were included in our systematic review. The enrolled studies were published between 2000 and 2024, and the sample size of the included studies ranged from 8 to 96. All 33 studies were a double-blind, randomized parallel-controlled study by design, and the remaining four were a randomized, single-blinded, crossover design. The average age of the participants ranged from 54.3 to 74.8, and the average duration of disease ranged from 2.5 to 16.5 years, with treatment duration ranging from 1 day to 3 months. six studies used intermittent theta-burst stimulation (iTBS) or continuous theta-burst stimulation (cTBS); the other used conventional rTMS as the intervention. The main outcome measures included UPDRS-III, FOG-Q, TUG, BDI, and MoCA: 38 studies used UPDRS-III as the outcome measure; Eleven studies used TUG; Eight studies used FOG-Q; Seven studies used BDI; Five studies used MoCA. The detailed characteristics of included studies were summarized in Table 1, Table 2.Table 1Study participant characteristics of the included trials.Included studiesCountryDesignAge(years)Disease duration(years)TreatmentdurationH & YInterventionsSimple sizeBenninger et al (2011)SwitzerlandParallelEG: 62.1 ± 6.9CG: 65.6 ± 9.0EG: 10.8 ± 7.1CG: 6.5 ± 3.42 weeks2– 4EG: real-iTMSCG: sham-iTMSEG: 13 CG: 13Benninger et al (2012)SwitzerlandParallelEG: 64.5 ± 12.5CG: 55.8 ± 9.1EG: 8.6 ± 4.1CG: 9.3 ± 6.82 weeks2–4EG: real-rTMSCG: sham-rTMSEG: 13 CG: 13Brusa et al (2006)ItalyCrossover61 ± 8.0416.4 ± 5.45 days1–3EG: real-rTMSCG: sham-rTMSEG: 10 CG: 10Chang et al (2016)KoreaCrossover71.9 ± 7.84.3 ± 1.85 daysUKEG: real-rTMSCG: sham-rTMSEG: 8 CG: 8Chung et al (2020)ChinaParallelEG1: 62.1 ± 5.7 EG2: 62.7 ± 6.8CG: 62.1 ± 5.7EG1: 7.5 ± 4.9 EG2: 5.2 ± 3.4 CG: 6.9 ± 3.33 weeksEG1: 2.2 ± 0.4 EG2: 2.2 ± 0.3 CG: 2.3 ± 0.3EG: real-rTMSCG: sham-rTMSEG1: 17 EG2: 17 CG: 16Cohen et al (2018)IsraelParallelEG: 64.4 ± 6.8CG: 66.8 ± 8.1EG: 4.7 ± 3.4CG: 5.6 ± 3.73 months2.0 ± 0.37EG: real-rTMSCG: sham-rTMSEG: 21 CG: 21Del Olmo et al (2007)SpainParallel61.7 ± 5.228.0 ± 5.010 days2–3EG: real-rTMSCG: sham-rTMSEG: 8 CG: 5Eggers et al (2015)GermanyCrossover65 ± 55.8 ± 5.01 day1.8 ± 0.8EG: real-rTMSCG: sham-rTMSEG: 13 CG: 13Filipović et al (2010)EnglandCrossover64.5 ± 9.615.6 ± 5.74 days3.3 ± 0.7EG: real-rTMSCG: sham-rTMSEG: 10 CG: 10Flamez et al (2016)BelgiumCrossover68.8 ± 10.314 ± 55 days3 ± 1EG: real-rTMSCG: sham-rTMSEG: 6 CG: 6Grobe-Einsler et al (2024)GermanyParallelEG: 66.06 ± 9.70CG: 70.41 ± 10.37UK5 days2EG: real-rTMSCG: sham-rTMSEG: 17 CG: 13Hamada et al et al (2009)JapanParallelEG: 65.3 ± 8.9CG: 67.4 ± 8.5EG: 8.1 ± 4.2CG: 7.8 ± 6.78 weeks2–4EG: real-rTMSCG: sham-rTMSEG: 55 CG: 43Ji et al (2020)ChinaParallelEG: 61.7 ± 1.57CG: 60.2 ± 1.97EG: 4.3 ± 0.52CG: 5.3 ± 0.8314 daysEG: 1.6 ± 0.12 CG: 1.7 ± 0.11EG: real-rTMSCG: sham-rTMSEG: 22 CG: 20Khedr et al (2003)EgyptParallelEG: 57.8 ± 9.2CG: 57.5 ± 8.4EG: 3.45 ± 2.3CG: 3.05 ± 2.110 days2–3EG: real-rTMSCG: sham-rTMSEG: 19 CG: 17Khedr et al (2006)EgyptParallelEG: 60.2 ± 9.48CG: 60.6 ± 10.6EG: 3.5 ± 0.7CG: 3.8 ± 0.96 days3–5EG: real-rTMSCG: sham-rTMSEG: 10 CG: 10Khedr et al (2019)EgyptParallelEG: 60.7 ± 8.8 CG: 57.4 ± 10EG: 5.7 ± 3.9 CG: 6.5 ± 3.710 daysEG: 3.1 ± 1.1 CG: 3.5 ± 1.0EG: real-rTMSCG: sham-rTMSEG: 19 CG: 11Khedr et al (2024)EgyptParallelEG: 61.82 ± 3.48 CG: 60.21 ± 1.64EG: 7.12 ± 3.48 CG: 5.87 ± 4.0810 daysUKEG: real-rTMSCG: sham-rTMSEG: 16 CG: 8Kim et al (2015)KoreaCrossover64.5 ± 8.47.8 ± 4.95 days3.0 ± 0.5EG: real-rTMSCG: sham-rTMSEG: 17 CG: 17Koch et al (2005)ItalyCrossoverEG: 60.75 ± 9.84EG: 16.5 ± 5.931 dayUKEG: real-rTMSCG: sham-rTMSEG: 8 CG: 8Lee et al (2014)KoreaCrossover71.6 ± 8.64.7 ± 2.61 day3.4 ± 0.5EG: real-rTMSCG: sham-rTMSEG: 20 CG: 20Li et al (2020)ChinaParallelEG: 61.7 ± 6.9 CG: 61.5 ± 8.4EG: 5.5 ± 3.7 CG: 6.5 ± 5.15 daysEG: 1.9 ± 0.6 CG: 1.8 ± 0.6EG: real-rTMSCG: sham-rTMSEG: 24 CG: 24Lomarev et al (2006)USAParallelEG: 63 ± 10 CG: 66 ± 10EG: 13.8 ± 6.8 CG: 10.8 ± 3.18 days2–4EG: real-rTMSCG: sham-rTMSEG: 9 CG: 9Maruo et al (2013)JapanCrossover63.0 ± 11.312.0 ± 6.33 days3.1 ± 0.5EG: real-rTMSCG: sham-rTMSEG: 21 CG: 21Mi et al (2020)ChinaParallelEG: 62.7 ± 10.6 CG: 65.6 ± 8.7EG: 9.2 ± 5.8 CG: 7.4 ± 4.810 daysEG: 2.6 ± 0.9 CG: 2.4 ± 0.9EG: real-rTMSCG: sham-rTMSEG: 20 CG: 10Brys et al (2016)CanadaParallelEG1: 64.9 ± 8.0 EG2: 59.6 ± 12.6 EG3: 64.6 ± 12.3 CG: 64.0 ± 7.4EG1: 7.3 ± 5.6 EG2: 8.4 ± 5.2 EG3: 7.7 ± 4.2 CG: 4.5 ± 2.210 daysEG1: 2.5 EG2: 2.92 EG3: 2.65 CG: 2.53EG: real-rTMSCG: sham-rTMSEG1: 20 EG2:14 EG3: 12 CG: 15Romero et al (2024)SpainParallelEG: 64.4 ± 6.38 CG: 66.89 ± 9.07EG: 6.0 ± 3.06 CG: 6.22 ± 4.1210 days1–3EG: real-rTMSCG: sham-rTMSEG: 10 CG: 9Pal et al (2010)HungaryParallelEG: 68.5 CG: 67.5EG: 6.0 ± 1.6 CG: 6.5 ± 1.710 daysUKEG: real-rTMSCG: sham-rTMSEG: 12 CG: 10Shimamoto et al (2001)JapanParallelEG: 65.1 ± 8.0 CG: 64.5 ± 6.6EG: 7.0 ± 4.2 CG: 7.3 ± 6.52 months1–4EG: real-rTMSCG: sham-rTMSEG: 9 CG: 9Shin et al (2016)KoreaParallelEG: 69 ± 6.75 CG: 67 ± 6.25UK2 weeks1–3EG: real-rTMSCG: sham-rTMSEG: 10 CG: 8Shirota et al (2013)JapanParallelEG1: 68.8 ± 7.6 EG2: 67.9 ± 8.4 CG: 65.7 ± 8.5EG1: 8.5 ± 7.3 EG2: 7.8 ± 6.6 CG: 7.6 ± 4.48 days2–3EG: real-rTMSCG: sham-rTMSEG1: 34 EG2:34 CG: 34Song et al (2024)ChinaParallelEG: 67.36 ± 6.99 CG: 70.50 ± 6.76EG: 6.18 ± 1.67 CG: 6.77 ± 2.0210 days2–4EG: real-rTMSCG: sham-rTMSEG: 22 CG: 22Spagnolo et al (2021)ItalyParallelEG1: 60.4 ± 8.1 EG2: 63.9 ± 10 CG: 64.2 ± 5.5EG1: 5.8 ± 2.1 EG2: 7.6 ± 4.9 CG: 7.2 ± 3.012 days2EG: real-rTMSCG: sham-rTMSEG1: 20 EG2: 19 CG: 20Sun et al (2024)ChinaParallelEG: 61.91 ± 1.73 CG: 59.41 ± 8.91EG: 4.25 ± 0.61 CG: 5.47 ± 3.9714 daysEG: 1.55 ± 0.11 CG: 1.68 ± 0.50EG: real-rTMSCG: sham-rTMSEG: 22 CG: 17Wen et al (2022)ChinaParallelEG: 67.54 ± 1.95 CG: 67.18 ± 2.86EG: 5.02 ± 0.99 CG: 4.13 ± 0.7510 daysEG: 2.42 ± 0.14 CG: 2.46 ± 0.23EG: real-rTMSCG: sham-rTMSEG: 13 CG: 11Yang et al (2012)ChinaParallelEG: 65.20 ± 11.08 CG: 67.00 ± 13.21EG: 6.40 ± 2.76 CG: 6.35 ± 3.5812 daysEG: 2.3 ± 0.4 CG: 2.4 ± 0.4EG: real-rTMSCG: sham-rTMSEG: 10 CG: 10Yokoe et al (2017)JapanCrossover69.1 ± 8.49.5 ± 3.23 days3.5 ± 0.6EG: real-rTMSCG: sham-rTMSEG1: 19 EG2: 19 EG3: 19 CG: 19Zhuang et al (2020)ChinaParallelEG: 60.58 ± 9.21 CG: 61.57 ± 13.25EG: 5.86 ± 4.36 CG: 5.71 ± 3.7710 daysEG: 2 ± 0.74 CG: 2.25 ± 0.93EG: real-rTMSCG: sham-rTMSEG: 19 CG: 14Makkos et al (2016)HungaryParallelEG: 67 ± 9.63 CG: 66 ± 5.93EG: 6 ± 5.19 CG: 5 ± 4.4410 days1–4EG: real-rTMSCG: sham-rTMSEG: 23 CG: 21Ma et al (2019)ChinaParallelEG: 59.94 ± 9.16 CG: 66.00 ± 8.55EG: 8.94 ± 5.48 CG: 7.50 ± 4.7210 daysEG: 2.42 CG: 2.4EG: real-rTMSCG: sham-rTMSEG: 18 CG: 10Arias et al (2010)SpainParallelUKUK10 days2–4EG: real-rTMSCG: sham-rTMSEG: 9 CG: 9Lench et al (2021)USAParallelEG: 66.6 ± 7.5 CG: 64.5 ± 8.9EG: 8.7 ± 7.1 CG: 8.0 ± 5.610 daysEG: 2.3 ± 0.4 CG: 2.3 ± 0.3EG: real-rTMSCG: sham-rTMSEG: 12 CG: 8W.He et al (2021)ChinaParallelEG: 70.0 ± 6.3 CG: 74.8 ± 6.9EG: 2.7 ± 1.5 CG: 2.5 ± 1.110 daysEG: 2.1 ± 1.1 CG: 2.5 ± 1.0EG: real-rTMSCG: sham-rTMSEG: 20 CG: 15Cheng et al (2022)ChinaParallelEG: 71.6 ± 5.1 CG: 73.9 ± 6.9UK14 daysEG: 3.0 ± 1.2 CG: 2.5 ± 1.0EG: real-rTMSCG: sham-rTMSEG: 11 CG: 16Lefaucheur et al (2004)FranceCrossover64 ± 211 ± 11 day3.4 ± 0.2EG: real-rTMSCG: sham-rTMSEG: 12 CG: 12Siebner et al (2000)GermanyCrossover57 ± 115.5 ± 3.41 day1–2EG: real-rTMSCG: sham-rTMSEG: 10 CG: 10EG, experimental Croup; CG, control Croup; USA, the United States of America; rTMS, repetitive transcranial magnetic stimulation; iTBS, intermittent theta burst stimulation; UK, unknown.Table 2Characteristics of included TMS variables.Included studiesrTMS siterTMS frequency (Hz)No. of pulsesessionIntensity,%On/off (evaluation)SessionOutcomesPost-evaluationBenninger et al (2011)M1&DLPFCiTBS (50)3880 %AMTOn & Off8Gait; UPDRS-III; Bradykinesia; BDI15 days; 45 daysBenninger et al (2012)M150UK80 %AMTOn & Off8Gait; UPDRS-III; FOG-Q; Bradykinesia; BDI15 days; 45 daysBrusa et al (2006)SMA1UK90 % RMTOn5UPDRS-III15 days; 30 days; 45 days; 60 daysChang et al (2016)M1101000590 % RMTOn5TUG; UPDRS-III;; Turn time; FOG-Q5 days; 12 daysChung et al (2020)Bilateral M125 and 110001280 % RMTOn12TUG; UPDRS-III1 day; Post1m; Post3mCohen et al (2018)PFC&M1DTMS&109002480 % RMTOn24TUG; UPDRS-III; BDI90 daysDel Olmo et al (2007)DLPFC104501090 % RMTOn10Gait; UPDRS-III1 dayEggers et al (2015)SMAcTBSUK90 % AMTOn & Off1UPDRS-IIIUKFilipović et al (2010)M111800490 % RMTOn & Off4UPDRS-III1 dayFlamez et al (2016)M1110001090 % RMTOff10UPDRS-III1 week; 2 weeksGrobe-Einsler et al (2024)M148UK50 % AMTOn10UPDRS-III; TUG5 days; 30 daysHamada et al et al (2009)SMA510008110 % AMTOn8UPDRS-III12 weeksJi et al (2020)left SMAcTBS10001480 % AMTOn14UPDRS-III; TUG1 week; 2 weeksKhedr et al (2003)M15200010120 % MTOn10UPDRS-III; Time of the25-m walk10 days; Post1mKhedr et al (2006)M11030006100 % MTOff6Time of the 25-m walk6 daysKhedr et al (2019)M11020001090 % RMTOn10UPDRS III; Self-assessment scale10 days; Post1m; Post2m; Post3mKhedr et al (2024)M12020001080 % RMTOn10UPDRS-III; BDI; PDSS10 daysKim et al (2015)M1-LL101000590 % RMTOn5FOG-Q; TUG; UPDRS-III5 days; 12 daysKoch et al (2005)SMA5 & 1900190 % RMTOff1UPDRS-IIIUKLee et al (2014)M1&SMA&DLPFC1010001120 % RMTOn1FOG-Q; TUG; UPDRS-III1 dayLi et al (2020)M1202000580 % RMTOn5UPDRS-III1 weekLomarev et al (2006)M1&DLPFC253008100 % MTOn & Off8UPDRS-IIIPost1mMaruo et al (2013)Bilateral M11010003100 % RMTUK3UPDRS-III; the 10-m walking test3 daysMi et al (2020)SMA2010001090 % RMTOn10FOG-Q; UPDRS-III12 days; Post1mBrys et al (2016)DLPFC + M1&M1&DLPFC10100010UKOn10UPDRS-IIIPost1mRomero et al (2024)Bilateral M11010001080 % RMTOn10TUG; UPDRS-III14 days; 16 daysPal et al (2010)left DLPFC56001090 % RMTOn10TUG; UPDRS-III; BDI;11 days; 41 daysShimamoto et al (2001)M10.2608NKOn8UPDRS-IIIPost1m; Post2mShin et al (2016)left DLPFC56001090 % RMTOn10BDI2 weeks; 6 weeksShirota et al (2013)SMA1&1010008100 % RMTOn8UPDRS-III; HAM-D9 weeks; 21 weeksSong et al (2024)Bilateral M11010001090 % RMTOn10FOG-Q; TUG; HAM-D; UPDRS-III; SS-180 turning steps10 days; 40 daysSpagnolo et al (2021)M1; M1-PFCDTMS&10100010100 % RMTOn12UPDRS-III11 days; 12 days; 40 daysSun et al (2024)SMAcTBS18001480 % RMTOff14UPDRS-III; MoCA14 daysWen et al (2022)SMA10100010110 % RMTOff10UPDRS-IIIUKYang et al (2012)M15100012100 %RMTOn12UPDRS-III; TUG4 weeksYokoe et al (2017)EG1: M1 EG2: SMA EG3: DLPFC1010003100 %RMTOn3UPDRS-III; MoCA;Walk time3 daysZhuang et al (2020)DLPFC1120010110 % RMTOn10UPDRS-III; HRSD; NMSQ; MoCA10 days; Post1m; Post3m; Post6mMakkos et al (2016)M156001090 % RMTOn10UPDRS-III; MoCA; BDI; TUG11 days; 41 daysMa et al (2019)SMA1010001090 % RMTOff10FOG-Q; UPDRS-III1 day; 5 days; 10 days; 24 days; 38 daysArias et al (2010)M116001090 % RMTOn & Off10Motor sections of the UPDRS10 daysLench et al (2021)SMA1120010110 % RMTOn & Off10FOG-Q; UPDRS-III10 daysW.He et al (2021)Left DLPFCiTBSUK100 % RMTUK10MoCA10 days; Post3mCheng et al (2022)Left DLPFCiTBS600*14UKUK10MoCA15 days; Post3mLefaucheur et al (2004)M10.5&10260080 % RMTOn1UPDRS-III1 daySiebner et al (2000)M15225090 % RMTOn1UPDRS-III1 dayM1, primary motor cortex; DLPFC, dorsolateral prefrontal cortex; MT, motor threshold; Post1m, 1-month postintervention; Post3m, 3-month postintervention; Post6m, 6-month postintervention; UK, unknown; UPDRS: Unified Parkinson’s Disease Rating Scale; rTMS, repetitive transcranial magnetic stimulation; DTMS, repetitive deep transcranial magnetic stimulation; EG, experimental Croup; CG, control Croup; iTBS, intermittent theta burst stimulation; RMT, resting motor threshold; AMT, active motor threshold; HRSD, Hamilton Rating Scale for Depression; NMSQ, Nonmotor Symptom Questionnaire; FOG-Q, freezing of gait questionnaire; BDI, Beck Depression Inventory; MoCA, Montreal Cognitive Assessment; TUG, time up and go; SMA, supplementary motor area; PFC, prefrontal cortex; UPDRS III, Movement Disorder Society–Unified Parkinson’s Disease Rating Scale motor score Part III.
The methodological quality and risk of bias were assessed according to the Cochrane risk of bias tool (Revman5.4). The detailed evaluation of each study was presented in Fig. 2, Fig. 3. Overall, the risk of bias for the included trials was relatively low.Fig. 2Risk of bias graph.Fig. 3Risk of bias summary.
45 Eligible trials were enrolled for meta-analysis using a random-effects model, with UPDRS-III total scores, subscores of UPDRS-III, and MoCA scores as primary. Besides, BDI scores, FOG-Q scores, and TUG (expressed in walking time) as secondary indicators to evaluate the effectiveness of rTMS for PD, and adverse events (AE) as safety indicators
The UPDRS-III scores were analyzed in 35 included trials involving 1398 patients. High heterogeneity was detected (I^2^ = 75 %; Q test p < 0.00001), and a random-effect model was adopted for analysis. The UPDRS-III scores in patients with PD significantly reduced after rTMS therapy (SMD = − 0.74; 95 % CI = [−0.97, −0.51]; p < 0.00001) (Fig. 4).Fig. 4Pooled results of UPDRS-III scores.
Among the subscores of UPDRS-III, 8 studies related to bradykinesia included 357 patients; 6 studies related to rigidity included 253 patients; 6 studies related to tremor included 253 patients; 5 studies related to gait included 245 patients; 4 studies related to postural instability included 209 patients; 3 studies related to axial symptoms included 105 patients. Subgroup analysis with random-effects model showed that the rTMS therapy significantly decreased subscores of UPDRS-III at bradykinesia score (SMD: −0.28; 95 %CI = [−0.50, −0.06]; p = 0.01), rigidity score (SMD: −0.66; 95 %CI = [−0.93, −0.39]; p < 0.00001; heterogeneity test p < 0.00001; I^2^ = 89 %), tremor score (SMD: −0.41; 95 %CI = [−0.67, −0.15]; p = 0.002; heterogeneity test p < 0.00001; I^2^ = 88 %), gait score (SMD: −0.42; 95 %CI = [−0.68, −0.15]; p = 0.002), and axial symptoms score (SMD: −0.77; 95 %CI = [−1.17, −0.36]; p = 0.0002; heterogeneity test p = 0.08; I^2^ = 61 %). However, no statistically significant differences were observed between the real-rTMS, and sham-rTMS at postural instability score (SMD: −0.19; 95 %CI = [−0.47, 0.08]; p = 0.17; heterogeneity test p = 0.50; I^2^ = 0 %) (Fig. 5).Fig. 5Pooled results of subscores of UPDRS-III.
The MoCA scores were analyzed in five included trials involving 185 patients. A fix-effects model demonstrated that compared with the control group, rTMS-treatment significantly increased the MoCA scores (WMD = 4.13; 95 %CI = [3.16, 5.09]; p < 0.00001; heterogeneity test p = 0.69; I^2^ = 0 %) (Fig. 6).Fig. 6Pooled results of MoCA scores.
Among 43 studies, the BDI scores were evaluated in 7 studies of 124 patients in the intervention group and 110 in the control group. The result of the Cochrane Q-test demonstrated high homogeneity among studies (I^2^ = 78 %; Q test p < 0.0001). The forest map results significantly improved BDI scores between groups (SMD = − 0.82; 95 % CI = [−1.43, −0.21]; p = 0.008) (Fig. 7).Fig. 7Pooled results of BDI scores.
The effect on FOG-Q was reported in 9 studies included in the meta-analysis. Similarly, a random-effects model indicated that there was a significant reduction in FOG-Q in the rTMS group in comparison to the control group (SMD = − 0.49; 95 %CI = [−0.87, −0.12]; p = 0.01; heterogeneity test p = 0.002; I^2^ = 64 %) (Fig. 8).Fig. 8Pooled results of FOG-Q.
Eleven studies assessed TUG (expressed in walking time) and were enrolled in the quantitative analysis. Our study with fixed effects model found that real rTMS treatment had a significantly beneficial effect on TUG (expressed in walking time) compared to placebo-controlled treatment (WMD = − 0.97; 95 %CI = [−1.38, −0.57]; p < 0.00001) with low heterogeneity (I^2^ = 15 %) (Fig. 9).Fig. 9Pooled results of TUG.
We performed a site subgroup analysis for the UPDRS-III scores to investigate the impact of rTMS on the UPDRS-III scores at the stimulation site. Pooled analysis showed that the rTMS group significantly decreased the UPDRS-III scores (SMD = − 0.68; 95 % CI = [−0.93, −0.43]; p < 0.00001; heterogeneity test p < 0.00001; I^2^ = 76 %), compared with the control group. Subgroup analysis with random-effects model showed that the rTMS therapy significantly decreased UPDRS-III score at the SMA site (SMD: −1.43; 95 %CI = [−2.20, −0.66]; p = 0.0003), M1 site (SMD: −0.45; 95 %CI = [−0.63, −0.27]; p < 0.00001; heterogeneity test p = 0.39; I^2^ = 5 %), and DLPFC site (SMD: −0.61; 95 %CI = [−1.01, −0.22]; p = 0.002; heterogeneity test p = 0.07; I^2^ = 47 %). However, no statistically significant differences were observed between the intervention and control group at the M1 and DLPFC site (SMD: −0.27; 95 %CI = [−0.72, 0.18]; p = 0.24; heterogeneity test p = 0.27; I^2^ = 24 %) (Additional file Fig. S1A).
To investigate the impact of rTMS therapy on UPDRS-III score at various combinations of rTMS site and rTMS frequency, we performed interaction between rTMS site and rTMS frequency subgroup analysis for UPDRS-III scores. Pooled analysis showed that rTMS were significantly decreased UPDRS-III scores (SMD: −0.84; 95 % CI [−1.12, −0.55]; p < 0.00001; heterogeneity test p < 0.00001; I^2^ = 77 %), compared with the control group. Further subgroup analysis with random-effects model showed that the rTMS group significantly decreased UPDRS-III scores at HF over M1 site (SMD = − 0.53; 95 %CI = [−0.74, −0.32]; p < 0.00001), HF over SMA site (SMD: −1.65; 95 %CI = [−2.53, −0.78]; p < 0.00001; heterogeneity test p = 0.0006; I^2^ = 80 %), and HF over DLPFC site (SMD: −0.46; 95 %CI = [−0.81, −0.12]; p = 0.008). Whereas, comparisons between both groups showed no statistical differences for LF over M1 site (SMD = − 0.33; 95 %CI = [−0.82, 0.16]; p = 0.18), LF over SMA site (SMD = − 0.85; 95 %CI = [−2.10, 0.40]; p = 0.16; heterogeneity test p = 0.004; I^2^ = 82 %), and cTBS over SMA (SMD = − 2.67; 95 %CI = [−5.53, 0.19]; p = 0.07; heterogeneity test p < 0.00001; I^2^ = 96 %) (Additional file Fig. S1B).
In the included literature, the main stimulation frequencies included high-frequency rTMS and low-frequency rTMS. Pooled analysis showed that the rTMS group significantly decreased UPDRS-III score (SMD = − 0.69; 95 %CI = [−1.27, −0.10]; p < 0.00001; heterogeneity test p < 0.00001; I^2^ = 68 %), compared with the control group. Subgroup analysis with random effects model presented that the rTMS intervention significantly decreased UPDRS-III scores in HF (SMD = − 0.62; 95 %CI = [−0.98, −0.34]; p < 0.0001; heterogeneity test p < 0.00001; I^2^ = 68 %), LF (SMD = − 0.62; 95 %CI = [−1.19, −0.04]; p = 0.04), and DTM (SMD = − 0.68; 95 %CI = [−1.27, −0.10]; p = 0.02; heterogeneity test p = 0.08; I^2^ = 60 %) (Additional file Fig. S1C).
To investigate the effect of rTMS on the UPDRS-III scores at various combinations of rTMS site and rTMS frequency. We performed the interaction between the rTMS site and rTMS frequency subgroup analysis for BDI scores. Pooled analysis showed that the rTMS group significantly decreased BDI scores (SMD = − 0.85; 95 %CI = [−1.48, −0.22]; p < 0.00001), compared with the control group. Subgroup analysis with a random-effects model showed that the rTMS group significantly reduced BDI score at HF over M1 sites (SMD = 3.47; 95 %CI = [2.56, 4.38]; p < 0.00001). However, the analysis results were no statistically significant at HF over DLPFC sites (SMD = − 1.79; 95 %CI = [−5.13, 1.55]; p = 0.29; heterogeneity test p = 0.04; I^2^ = 77 %), and iTBS over DLPFC sites (SMD = − 1.17; 95 %CI = [−4.18, 1.84]; p = 0.45) (Additional file Fig. S2).
We performed the interaction between the rTMS site and rTMS frequency subgroup analysis for FOG-Q. Pooled analysis showed that the rTMS group significantly decreased FOG-Q (SMD = − 0.49; 95 %CI = [−0.87, −0.12]; p = 0.01), compared with the control group. Subgroup analysis with a fix-effects model showed that the rTMS group significantly reduced FOG-Q at HF over M1 sites (WMD = − 0.98; 95 %CI = [−1.79, −0.17]; p = 0.02; heterogeneity test p = 0.53; I^2^ = 0 %) (Additional file Fig. S3).
Nineteen studies measured the incidence of AE [22], [24], [25], [26], [31], [34], [38], [39], [40], [47], [48], [51], [54], [56], [57], [58], [59], [60], [66]. Of these, most studies did not observe any AEs. The detailed symptoms of AEs in each study are described in Table 3.Table 3Information of adverse effects after the rTMS intervention.Included studiesAdverse eventsBenninger et al (2011)No obvious adverse effectsBenninger et al (2012)No obvious adverse effectsBrusa et al (2006)No obvious adverse effectsCohen et al (2018)10 cases of headache; 4 cases of dizziness; pain in the head or neck during treatment with three patientsFilipović et al (2010)No obvious adverse effectsJi et al (2020)No obvious adverse effectsKhedr et al (2003)No obvious adverse effectsKhedr et al (2006)an occasional mild, transient headache in some patientsLi et al (2020)2 cases of headaches and dizziness; 3 cases of headache; 2 cases of dizziness; 2 cases of tinnitus; 1 case of transient aggravation of gait disturbancesLomarev et al (2006)2 cases of painMaruo et al (2013)No obvious adverse effectsPal et al (2010)2 cases of mild transient headacheShin et al (2016)1 case of headacheShirota et al (2013)2 cases of tinnitus and headacheSong et al (2024)1 case of mild dizziness;1 case of mild headacheSpagnolo et al (2021)1 case of headache and dizziness respectivelyZhuang et al (2020)2 cases of mild headacheLefaucheur et al (2004)No obvious adverse effectsSiebner et al (2000)No obvious adverse effects
PD is the second most common neurodegenerative condition, which is characterized by the deterioration of motor activities as well as several non-motor symptoms (e.g. depression, dementia, pain, sleep disturbances, etc.) [67], [68]. Pharmacological treatment of PD is effective but impeded by side effects. Nonpharmacologic treatment primarily includes physiotherapy and speech therapy with limited success rates [69]. Thus, there is a need for alternative treatment strategies for PD. Repetitive transcranial magnetic stimulation (rTMS) is an emerging treatment technique for noninvasive brain stimulation, which is used to modulate numerous neurological disorders including PD [70]. Currently, 45 RCTs have been reviewed to evaluate the efficacy and safety of rTMS in treating PD. The results showed that rTMS therapy significantly improved motor symptoms and cognitive function in PD patients, as was indicated by the reduced UPDRS-III total scores, subscores of UPDRS-III and increased MocA scores. Remarkably, no significant adverse effects were found in the attended studies, which suggested the safety of rTMS therapy for PD. In addition, further analysis might provide more valuable data for subsequent clinical studies.
Our primary concern is the motor dysfunction of patients with PD, presenting with bradykinesia, resting tremor, rigidity, and postural instability. Importantly, these features are due to the alterations at different stages within the brain. The main pathological change is the loss of dopamine-producing neurons in the midbrain [71], [72]. Correspondingly, rTMS on the motor cortex releases dopamine in the caudate and putamen corresponding to their cortico-striatal projections [73]. The therapeutic mechanism of rTMS on motor function in PD might involve the motor cortex-basal ganglia-thalamo-cortical [74]. Our meta-analysis showed that rTMS have remarkable therapeutic effects on the motor in PD. The results were consistent with a recent meta-analysis, which indicated that rTMS significantly improved motor function in PD [73]. Additionally, sources of heterogeneity across studies were explored by performing subgroup analysis in this meta-analysis. One crucial parameter to refine rTMS therapy for PD is stimulation frequency. HF-rTMS presumably increases the excitability of neurons in the stimulated cortex; LF-rTMS presumably suppresses it [75], [76]. The frequency subgroup analysis of the UPDRS-III scores with a random-effects model indicated that patients who received HF-rTMS intervention and LF-rTMS intervention showed decreased UPDRS-III scores compared with counterparts who underwent sham stimulation. Similarly, animal studies also have shown that HF-rTMS reduces the death of striatal dopaminergic neurons by increasing intracranial levels of brain-derived neurotrophic factor, which in turn reduces motor symptoms [77]. Typically, the opposite result was discovered or no significant result was observed in both HF‐rTMS and LF‐rTMS [78]. The reasons for these differences may be related to disease severity, comorbidities, treatment course, stimulation parameters, and the number of included studies. In addition, repetitive deep transcranial magnetic stimulation (rDTM) exhibited some benefit on UPDRS-III score in PD [31]. We, therefore, recommend that future studies testing rDTMS should be focused on more patients following a longer treatment period. In light of this, it still calls for conducting more clinical trials to further determine the therapeutic effectiveness of multiple stimulation frequencies of rTMS in PD therapy in the future.
Another important parameter is a target subregion because many brain regions are involved in PD. To date, most studies have targeted the primary motor cortex (M1) or prefrontal cortex, and the supplementary motor area (SMA) [79], [80]. In included studies, rTMS over the motor cortex and the prefrontal cortex has shown beneficial effects on motor, mood, and cognitive symptoms without serious adverse effects. We aimed to investigate the site-specific effects of HF‐rTMS applied to different cortical regions on motor function, intending to establish a consensus therapeutic protocol. Thus, the subgroup analysis for the rTMS site of the UPDRS-III score showed that rTMS therapy significantly decreased the UPDRS-III score at the SMA site, M1 site, and DLPFC site. However, no statistically significant differences were observed between the intervention and control groups at the M1 combined with DLPFC sites. Further, the combination of rTMS frequency and rTMS site revealed that HF‐rTMS targeting the SMA, M1, and DLPFC showed significant effect sizes in motor symptoms. Whereas, LF‐rTMS over the M1 site, SMA site, cTBS over the SMA site, and HF‐rTMS over M1 + DLPFC were insignificant. Consistently, some studies have reported the effectiveness of HF-rTMS over M1 or SMA on motor function [22], [58], [81]. Importantly, our study found that HF-rTMS over DLPFC demonstrated improvement in motor symptoms in PD, which was probably caused by the stimulation effects spreading to distant cortical and subcortical regions via certain neural connections [82]. Typically, cTBS at the SMA stimulation model may be effective than sham stimulation in improving motor symptoms in PD. This conclusion should be interpreted cautiously and more clinical research is required to establish an optimized stimulation model in PD treatment.
In our study, the reduction of UPDRS III scores was mainly ascribed by improvements in several domains such as bradykinesia, rigidity, tremor, and gait, whereas postural stability remained unchanged. The results of TMS on postural stability are inconclusive. On the other hand, our results are consistent with a recent meta-analysis that demonstrated no improvement in postural stability following rTMS activation of M1 or the supplementary motor region [83]. However, the development of a proper protocol for PD treatment using rTMS requires future studies conducted on a larger scale.
Freezing of gait (FOG), characterized by sudden and brief episodes of inability to start the effective forward progression of the feet, is a common and severely disabling gait disorder in patients with advanced PD [84]. FOG usually occurs when initiating, turning, maneuvering through tight places, and getting close to a target, which is significantly impacted by several elements, including emotional cognition and environmental influences [85], [86], [87]. The detailed mechanism of FOG is not clear and there are limited therapeutic alternatives, as evidence for rehabilitation schemes, deep brain stimulation, and pharmacological treatment is inconclusive [88]. Studies have shown TMS might promote neuroplasticity by enhancing cortical functional connectivity [89]. Specifically, TMS stimulation may improve brain motor control and gait performance by reintegrating functional networks between motor-related brain areas [46]. The present meta-analysis indicated that real-rTMS therapy significantly reduced FOG-Q scores between groups, demonstrating that rTMS might serve as a therapy for improving FOG in PD patients. Additionally, the result of subgroup analysis showed greater effect sizes for rTMS over M1, compared to placebo treatment. Arias's study found that rTMS over the M1 significantly improved gait performance [23]. Interestingly, the results of three studies reported that rTMS over the SMA could alleviate gait dysfunction in patients with PD [46], [49], [90]. The SMA is situated in front of the M1 leg area, and it is engaged before the start of movement and plays a role in several motor processes [91]. Therefore, SMA stimulation may be a more appropriate site in PD patients with FOG. In addition, Lee et al. demonstrated that there was a therapeutic effect of 10 Hz rTMS targeting the DLPFC on FOG and gait function. Hence, high-quality, large-sample clinical studies should be conducted to further validate the specific rTMS parameters for PD.
Cognitive dysfunction is associated with abnormal neural activity in higher cognitive regions in PD. rTMS can regulate cortical excitability and improve local blood flow in the brain, which provides a theoretical basis for the treatment of cognitive impairment with rTMS [92]. It has been demonstrated that high-frequency rTMS in cognitive-related cortex improves the function of several cognitive domains, and the mechanism may be related to the increase of neuronal excitability, the driving of neuronal oscillatory electrical activity, and the increase of neuronal synaptic plasticity [93]. In the current meta-analysis, our results showed that rTMS significantly improved MocA scores, indicating that rTMS produces a beneficial effect on cognition. In the included studies, one study reported the positive effect of HF-rTMS over left DLPFC on cognitive function [29]. W. He et al. noted that iTBS over left DLPFC has a positive effect on improvement of cognitive function [62]. According to neuroimaging evidence, left DLPFC is associated with cognition control executive function, and attention [94]. Consequently, the DLPFC might be the prime target of rTMS for the treatment of cognition. Additionally, rTMS may have directly modulated the executive function center of PD by contributing to the improvement of selective domains in MoCA scores [66]. Unfortunately, due to the small sample, this conclusion should be interpreted cautiously. Therefore, more research must be performed to further confirm our findings, investigate the biological mechanisms, and establish optimal methodological features (such as targets, and parameters) for individuals with PD.
Depression is one of the most common non-motor features in patients with PD [95], [96]. Treatment protocols for depression have not been established in PD patients [97]. Functional brain connection anomalies, such as dysregulation between the emotion regulation network and the default mode network, are frequently observed in depressed patients. rTMS stimulation may help to modulate functional connectivity within cortical networks and enhance the brain’s ability to regulate emotions, thereby improving the patient’s overall mental health [89], [98]. Our meta-analysis showed that rTMS therapy significantly decreased BDI scores, compared with the control group. Further, we assess the site-specific effect of high-frequency rTMS applied over different cortical regions on depression for patients with PD. Consequently, the subgroup analysis for the rTMS site of the BDI score indicated HF-rTMS over the bilateral M1 was a beneficial treatment for depressive symptoms in patients with PD. In addition, HF-rTMS over left DLPFC also shows a better therapeutic effect than sham stimulation. HF-rTMS applied to the left DLPFC has effects similar to those of antidepressants [99]. Consistently, studies comparing the effects of fluoxetine and HF-rTMS on depression in PD patients found that both groups significantly improved on depression rating scales, such as the BDI and HRS [100], [101], [102]. Thus, HF-rTMS over left DLPFC or the bilateral M1 is an effective target for treating depression in PD. Lastly, we aimed to identify the most effective protocol among various kinds of methods. Thus, new stimulation protocols should be carried out in depressive patients with PD.
Based on the recently updated RCTs, we elevated the efficacy and safety of rTMS in the therapy of PD and offered new evidence for clinical application. However, there are still several certain limitations. Firstly, we assessed and examined the heterogeneity of enrolled studies and discovered a high degree of heterogeneity in UPDRS-III scores, indicating different severity and duration of PD among the enrolled patients. Secondly, the subgroup analysis was carried out to identify the potential factors that might affect rTMS in the therapy of PD (such as stimulation frequency, and stimulation target). The subgroup analysis produced preliminary findings, such as rTMS could reduce UPDRS-III total scores, subscores of UPDRS-III, FOG scores, and BDI scores, and increase MocA scores. Consequently, more strict assessments of rTMS treatment in PD are required in the future. Last but not least, the pooled results could be skewed because the sample size of this study was modest. Therefore, further multicenter, large sample, double-blind studies need to be carried out to address pressing concerns such as the suitable stimulation frequency, stimulation subregion, and total pulses in rTMS therapy and explore the long-term implications of rTMS on individuals with PD.
In conclusion, rTMS is safe and effective for treating motor symptoms, cognition, and depression of PD. However, it is urgent to establish a possible therapeutic protocol to fully maximize the potential of rTMS, which involves the optimization of stimulation frequency, stimulation site, and total pulses. In terms of motor functions, HF-rTMS over M1, SMA, or DLPFC has proved to be effective. For cognitive function, HF-rTMS over left DLPFC or iTBS over left DLPFC might be a beneficial effect. Collectively, it was highlighted that rTMS appears to be a promising therapeutic avenue across multiple domains, such as motor features and no-motor symptoms. Pathophysiology and mechanisms of rTMS treatment for PD remain incompletely understood. Thus, further research is still required to further expand our knowledge of rTMS and improve the therapeutic effects on patients with PD.
Not applicable.
Additional data used to support this study are presented in the Supplementary Material, further inquiries can be directed to the corresponding author.
Wenming Lu: Writing – review & editing, Writing – original draft, Software, Methodology, Data curation. Longxiang Yan: Writing – review & editing, Formal analysis, Data curation. Chuangguo Li: Visualization, Data curation. Kai Wang: Software, Formal analysis, Data curation. Qijing Wang: Supervision, Methodology. Sisi Xu: Visualization, Supervision, Conceptualization. Benguo Wang: Writing – review & editing, Writing – original draft, Supervision, Resources, Project administration, Methodology, Formal analysis.
Not applicable.
The authors are grateful for the financial support received from the Shenzhen Longgang District Science and Technology Innovation Bureau (LGKCYLWS2023004) and the Shenzhen Basic Research Project of Natural Science Foundation (JCYJ20250604180120026). The funding bodies played no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.