Introduction

Spinal cord injury (SCI) refers to damage and the severe loss of neurons in the spinal cord [1], resulting in sensory, motor, or autonomic dysfunction [2]. Since SCI is incomplete in most cases, increasing the connectivity of the descending corticospinal pathway [3] and the neuroplasticity of neurons in the motor cortex to the spinal cord can benefit the restoration of motor function [4]. However, spontaneous recovery after spinal cord injury is variable and usually unsatisfactory. Standard pharmacological and rehabilitative approaches have been reported to promote the recovery of motor function, but the overall effects remain limited and vary largely among individuals [5]. Some new approaches, such as stem cell transplantation [6] and exosome therapy [7], have been reported as having certain restorative effects on the nerves, but most such therapies are invasive or are in the animal testing stage, limiting their clinical application.

In recent years, non-invasive brain stimulation (NIBS), including repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS), has received extensive attention. NIBS mainly regulates the excitability of the cerebral cortex through electric or magnetic fields, which has the potential to be a non-invasive and relatively simple means of treatment [8]. In rTMS, the time-varying magnetic field acts on the cerebral cortex to produce induced currents, affecting the brain’s metabolism in specific brain networks. In general, high-frequency stimulation (5 Hz or higher) increases cortical excitability, while low-frequency stimulation (1 Hz or below) decreases cortical excitability [9]. Moreover, a patterned form of rTMS called theta-burst stimulation (TBS) consisting of three pulses at 50 Hz and repeated at 5 Hz to reach a total number of 600 pulses, has also been extensively used [9]. Furthermore, tDCS uses a weak direct current to modulate the activity of neurons in the cerebral cortex [8]. Anodal tDCS increases the excitability of the cortex, while cathodal tDCS decreases it [9].

There has been increasing interest in investigating the potential of NIBS in improving motor function after SCI. Several reviews on the effects of NIBS in SCI have indicated that the presumed neural excitability modulation comprising these two NIBS techniques may effectively improve motor function. Despite some studies that have shown the positive effects of NIBS on motor function after SCI [10, 11], inconsistent results persist [11,12,13,14]. In addition, several high-quality randomized controlled trials (RCTs) have been published in the last few years [10, 15]. The transient effect of single-session NIBS suggests that multiple sessions may be needed to achieve persistent effects [16, 17]. Therefore, this review aims to quantitatively investigate the effects of and update the evidence regarding NIBS on motor dysfunction by evaluating all existing published RCTs with multiple sessions involving people with SCI.

Methods

A preplanned protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO, CRD42016050444) by the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions [18]. Furthermore, our review is described based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [19].

Search strategy

Randomized control trials were identified by searching MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and the clinical trials registry and database of the U.S. National Institutes of Health (ClinicalTrials.gov) on December 10, 2021. We put no restrictions on the year of publication in our search. Only studies published in English with the full text available were included. The following broad search terms were used: “spinal cord injuries”, “non-invasive brain stimulation”, “NIBS”, “transcranial direct current stimulation”, “tDCS”, “transcranial magnetic stimulation”,” TMS” and a string of predetermined words, which yielded a high sensibility in the search for randomized controlled trials. Search strategies were developed for each database using both free-text terms and the controlled vocabulary (MeSH and Emtree). The PubMed search strategy is illustrated in Additional file 1. A manual search was also conducted from the reference lists of previous systematic reviews to identify additional relevant studies.

Selection criteria

The study participants included patients with motor dysfunction after SCI, meeting the diagnostic criteria for spinal cord injury (International standards for neurological classification of spinal cord injury, revised 2019). The intervention included NIBS (minimum 3 sessions with stimulation), including tDCS and rTMS. The sham NIBS groups were used as a comparator to evaluate the NIBS effects on recovering motor function. Studies were excluded if they were reviews or commentaries, basic experiments, a summary of meetings, book chapters, case reports, full text is not available, unpublished, or duplicate literature, including a sample with mixed neurologic conditions (neuropathic pain, neurogenic bladder, and so on).

Study selection

First, two reviewers (JMC and XLL) independently screened all records based on the titles and abstracts. Second, the full texts of the primarily selected studies was subsequently retrieved and further examined carefully. All duplicate documents were removed by using Endnote (version X8). The full text of all relevant studies was subsequently retrieved and further examined carefully. The reviewers attempted consensus to establish which studies fulfilled the eligibility criteria. Any disagreements were resolved by discussion with a third senior reviewer (YY).

Data extraction

Two reviewers independently conducted data extraction using a predefined data extraction form. Disagreements were resolved through discussion or, if required, adjudication by a third reviewer. The following variables were extracted from studies: (1) the general characteristics including authors, year of publication, (2) study designs, (3) sample characteristics including sample size, age, duration, SCI degree, level, etc., (4) interventions and control protocol type, (5) outcomes of motor function, (6) adverse effect. The mean scores and SD of the outcomes before and after the interventions were extracted, as well as the mean change scores and SD for meta-analyses. If the data reported in articles could not be used for data pooling, the authors of the articles were contacted for requesting the necessary data. Otherwise, the publications with unavailable data were removed.

Methodological quality assessment

The quality of the included studies was evaluated using RevMan software (version 5.4, Cochrane Collaboration, Oxford, United Kingdom). All included studies were evaluated by seven domains: (1) Random sequence generation (2) allocation concealment (3) blinding of participants (4) blinding of outcome assessment (5) inadequate outcome data (6) whether selecting to report outcome (7) other possible bias. According to Cochrane Collaboration’s tool for assessing the Risk of Bias [8, 18], each domain was classified as a high, low, or unclear risk of bias. Studies with a low risk of bias in three or more domains were suggested as trials of moderate to high methodological quality [8]. Usually, tests for funnel plot asymmetry are performed only when at least 10 studies are included in a meta-analysis [18]. Although 14 studies were included in this analysis, when sorted by outcomes, each outcome contained fewer than 10 studies. Thus, publication bias in these trials could not be assessed by graphical analysis of the funnel plot [18].

Outcome indicators

First, data were divided into several meta-analyses to identify possible NBIS effects. The primary outcomes were functional level, extremity strength, mobility, spasticity, and balance of each study (Table 1). When multiple outcome measures were reported without indication of a primary outcome, representative measures in the area of SCI research were chosen based on their validity and reliability [20]. We pooled the data using the change of the outcomes, if available. If not, they were estimated from the final and baseline values. Outcome measures chosen by these criteria are summarized in Table 1. The extremity strength was measured by the American Spinal Injuries Association (ASIA) impairment scale Upper Extremity Motor Score (UEMS) and Lower Extremity Motor Score (LEMS). The upper extremity function was measured by Jebsen Taylor hand function test (JTHFT). The mobility was measured as the 10 min walking test (10MWT), 6-min walking test (6MWT), and timed up and go test (TUG) respectively. The spasticity was measured using the upper/ lower Modified Ashworth Scale (MAS) and Hmax/Mmax amplitude ratio (H/M). Body balance was measured by Berg Balance Test (BBT). Subsequently, subgroup analyses based on the mean post-injury time were performed to identify the potential differences in primary outcome parameters between the subacute and chronic stages. Data from crossover studies were considered taking into account the two periods of the study to warrant a correct analysis of crossover studies and reduce bias [18].

Table 1 Sample characteristics, study design, and outcome measures used in the included NIBS studies
Full size table

Statistical analysis

Meta-analyses were performed using RevMan software (version 5.4, Cochrane Collaboration, Oxford, United Kingdom). To combine the outcomes of the included studies, the standardized mean difference (SMD) with 95% confidence intervals (CI) was calculated with a random-effect model and weighted by the pooled effect size. P < 0.05 was considered to be statistically significant. The statistical heterogeneity between the studies was assessed using Cochran’s Q test and quantified with the I2 statistic (I2 ≥ 50% indicated substantial heterogeneity) [21]. To identify the sources of heterogeneity, sensitivity analysis was conducted. Meta-regression was also performed to explore the influence of the mean age of participants on the effectiveness of NIBS interventions. Meta-regression and sensitivity analysis was carried out using STATA software (version 16.0).

Results

Identification of studies

A PRISMA flowchart of study selection by stage of the systematic review is shown in Fig. 1. A total of 14 studies met our inclusion criteria and were described in qualitative analysis. All studies were published in English. Four trials were conducted in Spain, 4 in the U.S.A, 1 in the UK, 1 in Finland, 1 in Denmark, and 2 in Austria.

Fig. 1
figure 1

Flow of studies through the review

Full size image

Study selection and characteristics

Characteristics of studies are summarized in Tables 1, and 2 and summarized below. All of the included controlled studies were RCTs. Eight trials compared parallel intervention groups [10,11,12,13, 15, 22,23,24] and 6 studies [25,26,27,28,29,30] used a crossover design. The test groups all received NIBS, and patients in the control group received sham stimulation. In addition, 10 of the 14 total studies included in these meta-analyses used other rehabilitation therapies combined with NIBS [10,11,12,13, 15, 22,23,24,25, 27]. The form and length of these therapies for motor function were both highly variable parameters. The length ranged from one week to three months. In terms of form, 2 studies combined physical therapy [16, 27], 5 combined robot training [11, 12, 15, 23, 24], 1 combined peripheral nerve stimulation [22], 1 combined massed practice [13],1 combined antispastic medication [30]. The mean post-injury time of enrolled participants in five studies was between 1 and 12 months (subacute stage) [10, 12, 23, 25, 30], and in nine studies was greater than 12 months (chronic stage) [11, 13, 15, 22, 24, 26,27,28,29]. As regards the stimulation pattern of TMS, all but one [22] of the remaining studies used excitability stimulation patterns. All articles involving tDCS used anodal stimulation. The frequency of treatment ranged from three [24, 27] to four [22], five [10,11,12,13, 15, 23, 26, 28,29,30], or seven [25] times per week. The duration of treatment ranges from 3 [25] to 36 [24] sessions. The treatment intensity (in terms of session duration) ranged from 200 s [26, 29] to 30 min [13] and the treatment time did not differ between the control and treatment groups. Due to the large heterogeneity in study designs about follow-up, the first assessment available after the intervention was chosen as a follow-up. The data were sorted and analyzed based on the outcome to provide an effective evaluation of NIBS on each of the aspects: 5 studies used results from UEMS [10, 12, 23, 26] on the strength of upper extremity, 2 used JTHFF on function of the upper extremity [11, 27] and 5 used LEMS [10, 12, 23, 25, 26] on the strength of lower extremity, 2 used UMAS [22, 26], 7 used LMAS [15, 22, 23, 25, 28,29,30] and 3 used H/M ratio [28,29,30] on spasticity, 3 used BBT [15, 24, 25] on body balance, 4 used 10MWT [10, 11, 15, 24, 25], 3 used 6MWT [10, 15, 24] and 4 used TUG [15, 24] on mobility. Thus, the results of these clinical trials were pooled in different meta-analyses.

Table 2 Detailed NIBS setting and number/frequency of sessions used in each included study
Full size table

A total of 225 subjects were included in the 14 studies. The basic characteristics of the included literature are shown in Table 1. The number of participants in each study ranged from 5 [22] to 43 [15]. The pooled sample was predominantly males (73.78%) with a mean (SD) of 44.31 (15.08) years of age and 2.77 (5.32) years of duration of post-injury. All included studies provided information on the level of spinal cord injury and baseline severity according to the ASIA. The number of patients with the injury level cervical was 128 (56.89%) of the sample, whereas thoracic was 92 (40.89%) and lumbar was 5 (2.22%). Complete SCI at A level of impairment in ASIA was present only in 1 patient, and incomplete SCI at B level was in 5(2.22%), level C in 115 (51.11%), and level D in 104 (46.22%) of these patients.

Adverse effects

Among 14 included studies, 8 reported no obvious adverse effects [12, 13, 22, 24,25,26, 28, 29]. One study has reported that 1 patient experienced a seizure during TMS stimulation [10]. Five studies reported minor adverse effects [11, 15, 23, 27, 30], such as tingling, itching, skin redness, sleepiness, facial muscle contraction, or headache, which were observed also in the sham group.

Quality

Figure 2 presents the authors’ judgments about the risk of each biased domain and percentages of risks across all included studies. 8 studies (57.14%) [12, 13, 22, 23, 28,29,30] reported adequate random sequence generation and 3 (21.43%) [10, 11, 15] hid the allocation scheme, all presented blinding of participants and personnel, 13 (92.86%) presented blinding of outcome assessment and all described a low risk for attrition, showing a low risk of bias. Therefore, all of the included studies presented moderate to high methodological quality.

Fig. 2
figure 2

Cochrane risk of bias assessment of the included studies. a Risk of bias graph; b Risk of bias summary

Full size image

Effects of interventions

Extremity strength and function

No greater improvements in the upper extremity performance (strength and function) assessed by UEMS and JTHFT were observed in the NIBS groups compared to the sham groups (SMD = 0.12, 95% CI = − 0.41–0.64; P = 0.65; I2 = 0%) (Fig. 3a). When only studies for the chronic stage were analyzed, the pooled effect remained insignificant (SMD = 0.12, 95%CI = − 0.41–0.64, P = 0.65, I2 = 0%) (Fig. 4a). In the NIBS group, lower extremity strength measured by LEMS was greater than that in the sham groups (SMD = 0.58, 95% CI = 0.02–1.14, P = 0.04, I2 = 52%) (Fig. 3b). When the studies for sub-acute stage or chronic stage were analyzed separately, the heterogeneities were both increased and the pooled effects were no longer significant (Subacute stage: SMD = 0.64, 95%CI = − 1.25–0.53, P = 0.16, I2 = 65%; Chronic stage: SMD = 0.51. 95% = − 0.42–1.45, P = 0.28. I2 = 62%) (Fig. 4b).

Fig. 3
figure 3

Weighted mean difference (95% CI) of the effect of NIBS compared with sham on (a) lower extremity strength by pooling data from 5 trials (LEMS), upper extremity strength by pooling data from 3 trials (UEMS) and upper extremity function by pooling data from 2 trails (JTHFT) in people with SCI; (b) Balance from 2 tails

Full size image
Fig. 4
figure 4

A subgroup analysis based on different SCI stages for the NIBS effects compared with sham on (a) upper extremity strength, (b) lower extremity strength, (c) LMAS, (d) H/M, (e) 6MWT and (10) TUG

Full size image

Balance

Significant improvements by BBT were observed in the NIBS groups compared to the sham groups (SMD = 0.64, 95% CI = 0.05–1.24, P = 0.03, I2 = 0%) (Fig. 3c).

Spasticity

The changes of spasticity in upper limbs assessed by UMAS were not more significant in the NIBS groups than in sham groups (SMD = − 1.25, 95% CI = − 2.83–0.34, P = 0.12, I2 = 62%) (Fig. 5a). The changes of spasticity in lower limbs assessed by LMAS were more significant in the NIBS group than those in the sham groups (SMD = − 0.61, 95% CI = − 1.20 to − 0.03, P = 0.04, I2 = 64%) (Fig. 5a). When only subacute or chronic stage studies were analyzed for LMAS, heterogeneities were slightly changed and the pooled effects were no longer significant (Subacute stage: SMD = -0.71, 95%CI = − 1.85–0.43, P = 0.22, I2 = 79%; Chronic stage: SMD = − 0.57, 95%CI = − 1.32–0.19, P = 0.14, I2 = 58%) (Fig. 4c). However, overall changes in spasticity of lower limber measured by H/M ratio were similar in the NIBS groups and sham groups (SMD = − 0.95, 95% CI = − 2.64–0.73, P = 0.27, I2 = 86%) (Fig. 5b). When only studies for chronic stage were analyzed, heterogeneity was further increased and the pooled effects were still insignificant (SMD = − 1.47, 95% CI = − 4.29–1.34, P = 0.30, I2 = 91%) (Fig. 4d).

Fig. 5
figure 5

Weighted mean difference (95% CI) of the effect of NIBS compared with sham on (a) upper extremity spasticity by pooling data from 2 trials (UMAS) and lower extremity strength by pooling data from 7 trials (LMAS) in people with SCI; (b) H/M ratio from3 tails

Full size image

Mobility

Mobility was similar in the NIBS groups and sham groups, which was evaluated by the gait distance of 6MWT (SMD = − 0.17, 95% CI = − 0.68–0.34; P = 0.51; I2 = 63%) (Fig. 6a), the speed of 10MWT (SMD = 0.85, 95% CI = − 0.07–1.76; P = 0.07; I2 = 14%) (Fig. 6b), time-to-complete the 10MWT (SMD = − 0.35, 95% CI = − 0.88–0.18; P = 0.19; I2 = 0%) (Fig. 6c), and TUG (SMD = 0.01, 95% CI = − 0.51–0.52, P = 0.98, I2 = 16%) (Fig. 6d). For 6MWT, when only studies for the chronic stage were analyzed, heterogeneity was further increased and the pooled effects were still insignificant (SMD = − 0.44, 95% CI = − 1.53–0.65, P = 0.37, I2 = 81%) (Fig. 4e). For TUG, the heterogeneity was decreased when only studies for the subacute stage were analyzed (I2 = 0%) and increased for only chronic stage studies (I2 = 71%). However, the pooled effects were still not significant (substage stage: SMD = − 0.07, 95% CI = − 0.72–0.59, P = 0.84; chronic stage: SMD = 1.43, 95% CI = − 2.40–5.26, P = 0.46) (Fig. 4f).

Fig. 6
figure 6

Weighted mean difference (95% CI) of the effect of NIBS compared with sham on mobility by pooling data from (a) 3 trials of 6MWT, (b) 2 trials of 10MWT in speed, (c) 2 trials of 10MWT in time- to-complete, (d) 4 trials of TUG

Full size image

Meta-regression

The age of participants is a highly heterogeneous parameter. The univariate meta-regression analysis was performed to identify any association between mean age and effect size. The results showed that the mean age was not a significant predictor of the effect size. However, these analyses may be underpowered, given the small number of studies involved (much less than 10) [31, 32]. The results of the univariate meta-regression are presented in Additional file 2.

Sensitivity analysis results

Sensitivity analysis revealed that the heterogeneity across subgroups did not change after excluding any one study, suggesting the source of the heterogeneity was multifaceted. The results are shown in Additional file 3.

Discussion

As a new neuromodulation technique, NIBS has been reviewed for its potential to improve motor function after SCI [3, 17]. The present systematic review and meta-analysis evaluate the effects and summarize the safety profiles of NIBS. The data for participants of the included trials in this review demonstrated evidence that NIBS has positive effects on the strength and spasticity of the lower extremities, as well as balance.

From a motor control perspective, damage to spinal tracts disturbs the information transmission from the brain to the spinal cord. In addition, the damage also results in maladaptive reorganization of the entire neuraxis, contributing to motor dysfunction [17]. At the spinal level, the maladaptive reorganization of spinal circuits leads to spasticity, due in part to the loss of descending control of inhibitory spinal circuits [33]. Clinically, injuries are divided into the categories of (neurologically) complete or incomplete, depending on the presence or absence of neurological functions below the segmental level of injury. Numerous histological analyses and electrophysiologic studies have demonstrated that most patients diagnosed with complete SCI with loss of all neurological functions below the injury have residual physiological or anatomical continuity of the central nervous system tracts across the lesion. These residual tracts provide a fertile ground for NIBS, which is concerned with establishing central axonal regeneration and reestablishing physiological reconnections [34].

NIBS is applied over the motor cortex in SCI patients to take advantage of neuroplasticity to activate the residual axon and establish functional connectivity in the corticospinal tract. Some clinical studies have indicated that the ultimate therapeutic effects are produced by affecting signaling in the nervous system; namely, by exciting, inhibiting, or regulating neuronal and neural network activities [35, 36]. In addition, several studies in animal models of SCI have also suggested the benefits of NIBS. Some studies have indicated that following injury, NIBS can enhance the spontaneous collateral or regenerative sprouting of corticospinal tracts, increase the regeneration rate of axons, as well as produce motor recovery corresponding to increased axonal growth [36, 37]. Another study showed that tDCS can increase the expression of a brain-derived neurotrophic factor in mouse cortical slices, which can promote changes in synaptic plasticity [38]. Poirrier et al. reported that after eight weeks of treatment with 10 Hz rTMS, a significant positive correlation between the final motor function of SCI in animal models and the grey matter density of the serotonergic fibers in the spinal segment [39]. Cao et al. reported that the ability of rTMS to alleviate spasticity and promote motor function following SCI might be related to the varying degrees of up-regulation of GABA receptors [40] and potassium-chloride cotransporter-2 protein [41]. However, the study by Poirrier et al. has suggested a mechanism by which rTMS is beneficial in low thoracic lesions because it activates the central locomotor generator [42].

In our meta-analysis, the motor cortex was selected as a stimulation site in all studies, and NIBS showed a significant effect on lower-extremity strength and balance. However, there was no greater improvement in functional mobility in the NIBS group compared to the sham group. In addition, high heterogeneity was observed in the LEMS and 6MWT. Some previous studies have shown that NIBS targeting the motor regions of the lower extremities can activate spinal circuits to improve walking function. However, the evidence of the value of NIBS in improving lower-extremity function in SCI remains limited [17]. In our meta-analysis, the presented results consist of previous studies wherein NIBS showed a significant effect on the lower-extremity strength and balance with the motor cortex selected as the stimulation site. Most of the participants in the analyses were older and in the chronic stage of SCI, which may be associated with a low improvement rate in terms of functional mobility [4]. On the other hand, the effects of NIBS on functional recovery after SCI also depend on the severity of the injury and individuals with a high severity of SCI may have poorer potential for neurological recovery [43]. In total, the fact that one participant with level A and five with level B according to the ASIA grade were included in the analyses may be also associated with poor general improvement in terms of functional mobility in the present study. Therefore, there was no greater improvement in functional mobility in the NIBS group compared to the sham group. In addition, benefits attained with NIBS alone might be confounded by the use of some add-on form of rehabilitation therapy in part [44]. These conventional therapies were not standardized among the studies included in the present review. We also did not account for the differences in form and duration of these add-on therapies in the calculations of efficacy. Hence, high heterogeneity was observed in the LEMS and 6MWT. Furthermore, the results of the mobility analyses may have been influenced by the small number of participants in the primary studies, which is associated with low statistical power and therefore a high probability of type II errors [45].

The present meta-analysis showed that NIBS had a significant benefit in addressing the spasticity of lower limbs evaluated by LMAS, which was consistent with the review performed by Korzhova et al. [46]. However, despite previous studies that have associated the spasticity of the lower extremities with the results of some neurophysiological examinations such as F-waves [22, 35], H reflex, and H/M ratio [22], the H/M ratio in our results failed to show significant change after the NIBS sessions. The results of the H/M ratio analysis may have been affected by the smaller number of studies (only three studies were included). Another equally important cause of the large heterogeneity in the effects of NIBS on spasticity, found in the studies, was the use of different stimulation protocols (frequency, total number of stimuli, stimulation intensity). For example, several TMS studies demonstrated the beneficial effects of theta bursts with a total frequency of 5 Hz in patients with SCI [29]. In other studies of TMS, the effect was also observed for high-frequency stimulation [28, 30]. Therefore, these differences suggest that the potential value of NIBS in spasticity post-SCI requires follow-up in additional in-depth studies.

Although a limited number of clinical trials have shown positive effects of NIBS on upper-extremity motor function in SCI patients, the present study did not find conclusive results of the UMAS, UEMS and JTHFT to support this idea. In line with this, our findings are consistent with the aforementioned studies by Lu et al. and Mateo et al. [47, 48]. The insufficient evidence could be explained by several factors. First, unlike motor dysfunction of the lower extremities, which can occur in the cervical, thoracic, and lumbar levels of SCI, motor dysfunction of the upper extremities only occurs in patients with cervical levels with relatively low incidence. Secord, most of the patients with cervical SCI have severe secondary complications, leading to poorer adherence to a trial’s training specifications [47, 48]. Third, unilateral hemisphere stimulation may not be the most efficacious approach for improving upper-extremity function in SCI, as the motor dysfunction of the upper extremities after SCI is typically bilateral[17]. Fourth, arm and hand function are a complex issue both in SCI and non-SCI patients with tetraplegia, including a wide variety of highly acyclic movements that cannot be easily objectively measured [47, 49].

It should be noted that the differences in demographic indices including age and post-injury time may influence the efficacy of rehabilitation treatments [43, 50, 51]. However, to our surprise, the meta-regression showed that the mean age was not a significant predictor of effect size for the outcome parameters in this review. The first probable explanation for this phenomenon might be the large age range of participants included in each trial. The second explanation may be the small number of studies involved (much less than 10) for each univariate meta-regression analysis, which could affect the results of the meta-regression. The influence of post-injury time on efficacy also has previously been widely investigated [52, 53]. However, the rather limited number of included studies and small sample size in the present study do not allow for firm conclusions to be made from such comparisons between subacute and chronic stages. Further studies directly comparing the different NIBS effects between populations for different ages and stages are necessary for the future.

This review found that some studies reported mild adverse events, such as headaches, facial muscle contraction, and tingling [13, 17, 25, 26, 29]. The most concerning adverse event was a seizure after rTMS [4]. Beyond that, there are no other types of major adverse events were observed in the current review, and no studies reported deterioration in motor function after NIBS. To establish the routine use of NIBS for SCI, it is necessary to develop a method to identify the lowest-risk stimulation parameters. Therefore, we suggest that more clinical evidence is needed in the future regarding the relationship between safety and stimulation parameters in order to improve the effectiveness of treatment.

Overall, our results are important for the emerging field of the use of NIBS in the motor recovery of lower limbs after SCI and support previous findings. Additionally, the present systematic review provides important information for future studies designed to address aspects of motor rehabilitation using NIBS as a rehabilitation tool for individuals after SCI.

Limitations

Some limitations in the present study should be noted. First, the methodological quality of a few included trials was low, and the study designs differed greatly. Eligibility criteria, random sequence generation, and allocation concealment were heterogeneous or not clearly stated in the articles. The power of the findings and their implications for clinical practice are thereby diminished. Second, our results are restricted to the short-term effects of NIBS, as the included studies did not assess long-term follow-up. Future original research should consider this aspect. Third, we only included articles published in English, which may cause bias if relevant studies have been published in other languages. Fourth, while tDCS and TMS are different types of stimulation with different working mechanisms, our findings indicate that they might trigger comparable effects on motor function. However, we were unable to perform sub-analyses to clarify different types of NIBS techniques and different stimulation parameters due to the small number of studies extracted and the variability in stimulation parameters reported, thus limiting our understanding of the positive changes in motor function promoted by NIBS. Fifth, functional neuroimaging and neurophysiological markers are needed to facilitate a more precise application of NIBS in SCI-related motor dysfunction. Sixth, the forms and parameters of combined rehabilitation therapies during NIBS also varied across the studies. Finally, it remains unclear whether the post-injury time, age, severity of the injury, lesion level, and type of injury are influential factors in NIBS results. These factors should be considered in the formation of homogeneous samples to determine whether these factors are predictors of better motor responses after NIBS.

Conclusion

From the concept of rehabilitation aimed at improving neuroplasticity, NIBS may be a promising complementary treatment when used in conjunction with conventional therapies or training to enhance motor function in patients with SCI. Our results provide initial evidence of the efficacy of NIBS in improving motor dysfunction in the lower extremities of SCI patients and encourage further high-quality research in this field.