Efficacy and Safety of Transcranial Direct Current Stimulation on Post-Stroke Dysphagia: A Systematic Review and Meta-Analysis

Dysphagia is one of the most common symptoms in patients after stroke onset, which has multiple unfavorable effects on quality of life and functional recovery. Transcranial direct current stimulation (tDCS) is a non-invasive brain stimulation that is widely used to improve deglutition function. Recently, some studies have confirmed that tDCS enhances deglutition function after stroke. However, the number of evaluation indexes used in those studies was small, and the number of trials included was limited. Most importantly, the optimal stimulation protocol is still uncertain and the safety of tDCS has not been reviewed. Therefore, we conducted a systematic review and meta-analysis to address these shortcomings. Methods: Seven databases were searched entirely, including Pubmed, Cochrane Library, Web of Science, China National Knowledge Infrastructure (CNKI), Chinese Biomedical Literature Service System (SinoMed), Wan-fang database, and the Chinese Scientific Journals Database (VIP) from inception to 31 December 2021. Two reviewers independently evaluated the eligibility of retrieved data according to the selection criteria and assessed the methodological quality of the studies using the Cochrane risk of bias tool. Outcomes, measures, and indicators used in this study included the dysphagia outcome and severity scale (DOSS), modified Mann assessment of swallowing ability (MMASA), functional oral intake scale (FOIS), functional dysphagia scale (FDS), and Kubota’s water-drinking test (KWDT). Sensitivity and subgroup analyses were performed to evaluate the intervention effect more specifically. Results: Fifteen trials with a total of 787 participants (394 subjects in the tDCS groups were treated with true tDCS, and 393 subjects in the control groups were wait-listed or treated with sham tDCS) involving tDCS for dysphagia after stroke and were included in the meta-analysis. Results of this meta-analysis confirmed that tDCS had a positive effect on post-stroke dysphagia. Subgroup analyses revealed that bilateral and high-intensity stimulation with tDCS had a more significant impact on post-stroke dysphagia. Furthermore, no adverse events occurred during the application of tDCS for post-stroke dysphagia. Conclusion: tDCS can promote the recovery of deglutition function in patients with dysphagia after stroke. In addition, bilateral stimulation and high-intensity stimulation may have better effects. However, the safety evidence for tDCS and post-stroke dysphagia is insufficient.


Introduction
Dysphagia is one of the most common problems following a stroke, with a high incidence of 80% [1]. Although the prevalence of dysphagia gradually decreases over time, 50% of patients still have symptoms of dysphagia at six months after stroke onset [2]. Dysphagia increases the incidence of undernutrition, dehydration, aspiration pneumonia,

Participants
A study was included if the adult participants (>18 years of age) were diagnosed with swallowing dysfunction after stroke and the stroke type was either cerebral hemorrhage or cerebral infarction. A study was excluded if dysphagia was caused by traumatic brain injury, oropharyngeal disease, esophageal disease, or mental disorders. In addition, dysphagia associated with neuromuscular disorders was also excluded.

Interventions
The intervention in the experiment group included tDCS alone or in combination with conventional therapy, and the control group included conventional treatment and/or sham tDCS.

Outcomes
The primary outcome indicator for this study was the dysphagia outcome and severity scale, and the secondary outcome indicators included the modified Mann assessment of swallowing ability, functional oral intake scale, functional dysphagia scale, and Kubota's water-drinking test.

Literature Selection and Data Extraction
One reviewer performed literature searches according to the specified search strategies and downloaded the related citations. All of the selected literature was imported into Endnote X9 (Clarivate, Philadelphia, PA, USA) and duplicate citations were removed using electronic/manual checking. Subsequently, two independent reviewers screened and identified the titles and abstracts of the remaining literature and then independently retrieved the literature that fulfilled the inclusion criteria. Discussion with the corresponding author resolved any inconsistent results between the reviewers. After the initial screenings, two reviewers independently extracted the relevant data from the identified studies. The following information was extracted from each study: general information (authors, publication year), demographic data (sample size, age, gender, stroke onset, stroke type, and stroke location), intervention (tDCS group, control group), tDCS protocol (site of stimulation, intensity of stimulation, duration of stimulation, treatment period), outcome measure (dysphagia outcome and severity scale, modified Mann assessment of swallowing ability, functional oral intake scale, functional dysphagia scale, and Kubota's water-drinking test), and adverse effects.

Data Analysis 2.7.1. Assessment of Risk of Bias in Included Studies
Two independent reviewers evaluated the risk of bias in each study by using the Cochrane risk of the bias assessment tool [20]. This assessment tool mainly includes seven domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other sources of bias. Each domain of the individual study was classified as having a high, low, or unclear risk. Any discordance that occurred between the two reviewers was resolved by discussions with the corresponding author.

Statistical Analysis
All data analyses in this study were conducted with R software (available at: http: //www.r-project.org/ (accessed on 10 January 2022), version 3.6.3). Continuous data were calculated as mean differences (MD) with 95% confidence intervals (CI). The I 2 statistic was used to evaluate the heterogeneity of the studies (with I 2 statistic > 50% indicating statistically significant heterogeneity). Fixed effects or random-effects models were used according to the heterogeneity (I 2 statistic > 50%, random effects models; I 2 statistic < 50%, fixed effects model). In addition, sensitivity analyses and subgroup analyses were carried out to dissect the heterogeneity. A forest plot was used to detect publication bias. A total of 255 published studies were identified (42 references from CNKI, 44 references from Wan-fang, 25 references from VIP, 26 references from SinoMed, 22 references from Pubmed, 44 references from Cochrane Library, 52 references from Web of Science) and imported into Endnote X9. After eliminating duplicates,101 articles remained. We then excluded reviews, case reports, and animal experiments, and 67 studies remained. Mixed interventions, and outcome indicators that did not include dysphagia outcome and severity scale, functional oral intake scale, modified Mann assessment of swallowing ability, functional dysphagia scale, or Kubota's water-drinking test were also excluded. Finally, 15 trials were considered after reading the full text. A detailed flowchart for the screening process is shown in Figure 1.

Statistical Analysis
All data analyses in this study were conducted with R software (available at: http://www.r-project.org/ (accessed on 10 January 2022), version 3.6.3). Continuous data were calculated as mean differences (MD) with 95% confidence intervals (CI). The I 2 statistic was used to evaluate the heterogeneity of the studies (with I 2 statistic > 50% indicating statistically significant heterogeneity). Fixed effects or random-effects models were used according to the heterogeneity (I 2 statistic > 50%, random effects models; I 2 statistic < 50%, fixed effects model). In addition, sensitivity analyses and subgroup analyses were carried out to dissect the heterogeneity. A forest plot was used to detect publication bias. A total of 255 published studies were identified (42 references from CNKI, 44 references from Wan-fang, 25 references from VIP, 26 references from SinoMed, 22 references from Pubmed, 44 references from Cochrane Library, 52 references from Web of Science) and imported into Endnote X9. After eliminating duplicates,101 articles remained. We then excluded reviews, case reports, and animal experiments, and 67 studies remained. Mixed interventions, and outcome indicators that did not include dysphagia outcome and severity scale, functional oral intake scale, modified Mann assessment of swallowing ability, functional dysphagia scale, or Kubota's water-drinking test were also excluded. Finally, 15 trials were considered after reading the full text. A detailed flowchart for the screening process is shown in Figure 1.

Characteristics of Included Studies
A total of 15 articles were included, consisting of 787 patients (393 patients in the control group and 394 patients in the tDCS group) with dysphagia after stroke. The interventions in the control group included CT only (n = 7), CT + sham tDCS (n = 8), and the intervention in the tDCS group was CT + tDCS. Of these, two studies did not report the course of stroke. The maximum current intensity for tDCS was 2 mA, and the minimum was 1 mA. The shortest treatment period for the intervention was five days, and the longest was two months. For outcome measure, five trials used the dysphagia outcome and severity scale, four trials used the modified Mann assessment of swallowing ability, four trials used the functional oral intake scale, three trials used the functional dysphagia scale, and two trials used the Kubota's water-drinking test. For adverse effects, six studies reported no adverse events, including skin redness, skin break, epilepsy, seizures, headaches, visual disturbances, skin irritation, or visual disturbance, and the remaining studies provided no information on adverse effects. The detailed characteristics of the included studies are shown in Table 1.  Figure 2 summarizes the risk of bias in the included studies. Six trials reported a method of random sequence generation and were assessed as low risk of bias [21][22][23][24][25][26]; two trials used the wrong randomization method and were evaluated as high risk of bias [27,28]; one trial described the methods of allocation concealment and was regarded as low risk of bias [23]; four trials mentioned the method for blinding the participants and personnel and were considered as low risk of bias [21,[29][30][31]; four trials mentioned the method of blinding the outcome assessment and were assessed as low risk of bias [30][31][32][33]; four trials had complete outcome data and were considered as low risk of bias [23,26,29,32]; one trial had incomplete outcome data and was regarded as high risk of bias [31]; one trial did not involve selective reporting and was assessed as low risk of bias [26]. In addition, all trials were not clear about other sources of bias.

Risk of Bias Assessment
ported no adverse events, including skin redness, skin break, epilepsy, seizures, headaches, visual disturbances, skin irritation, or visual disturbance, and the remaining studies provided no information on adverse effects. The detailed characteristics of the included studies are shown in Table 1. Figure 2 summarizes the risk of bias in the included studies. Six trials reported a method of random sequence generation and were assessed as low risk of bias [21][22][23][24][25][26]; two trials used the wrong randomization method and were evaluated as high risk of bias [27,28]; one trial described the methods of allocation concealment and was regarded as low risk of bias [23]; four trials mentioned the method for blinding the participants and personnel and were considered as low risk of bias [21,[29][30][31]; four trials mentioned the method of blinding the outcome assessment and were assessed as low risk of bias [30][31][32][33]; four trials had complete outcome data and were considered as low risk of bias [23,26,29,32]; one trial had incomplete outcome data and was regarded as high risk of bias [31]; one trial did not involve selective reporting and was assessed as low risk of bias [26]. In addition, all trials were not clear about other sources of bias.   Five RCTs involved the dysphagia outcome and severity scale [26,27,29,32,33]. After carefully reading the full text of the corresponding studies, the intervention protocols in the included five trials were different. Hence, a subgroup analysis was performed according to the intervention protocol. Since the I 2 statistic > 50%, a random-effects model was used to perform the meta-analysis. Treatment with tDCS compared with no tDCS showed a significant difference (MD = 1.26, 95% CI = 0.68; 1.84) and the corresponding result is shown in Figure 3A. A subgroup analysis showed that a high stimulation intensity (1.6-2 mA) had a larger positive effect on post-stroke dysphagia than a low stimulation intensity (1-1.5 mA). The corresponding results are shown in Figure 3B. In addition, a sensitivity analysis showed that the results of this meta-analysis were stable.

Risk of Bias Assessment
. Clin. Med. 2022, 11, x. https://doi.org/10.3390/xxxxx www.mdpi.com/journal/jcm a significant difference (MD = 1.26, 95% CI = 0.68; 1.84) and the corresponding result i shown in Figure 3A. A subgroup analysis showed that a high stimulation intensity (1.6-2 mA) had a larger positive effect on post-stroke dysphagia than a low stimulation intensity (1-1.5 mA). The corresponding results are shown in Figure 3B. In addition, a sensitivity analysis showed that the results of this meta-analysis were stable.

The Meta-Analysis Results for the Modified Mann Assessment of Swallowing Ability
A total of six studies used the modified Mann assessment of swallowing ability [21,22,24,25,27,28]. Since the I 2 statistic > 50%, a random-effects model was used to perform the meta-analysis. The results of this meta-analysis showed that when tDCS was com pared with no tDCS, there was a significant difference (MD = 7.57, 95% CI = 4.53; 10.62) The corresponding results are shown in Figure 4A. A subgroup analysis showed that bi lateral brain stimulation had a larger positive effect (MD = 6.19, 95% CI = 4.65; 7.74) than

The Meta-Analysis Results for the Modified Mann Assessment of Swallowing Ability
A total of six studies used the modified Mann assessment of swallowing ability [21,22,24,25,27,28]. Since the I 2 statistic > 50%, a random-effects model was used to perform the meta-analysis. The results of this meta-analysis showed that when tDCS was compared with no tDCS, there was a significant difference (MD = 7.57, 95% CI = 4.53; 10.62). The corresponding results are shown in Figure 4A. A subgroup analysis showed that bilateral brain stimulation had a larger positive effect (MD = 6.19, 95% CI = 4.65; 7.74) than undamaged brain stimulation (MD = 5.87, 95% CI = 2.40; 9.35). The corresponding results are presented in Figure 4B. In addition, a sensitivity analysis showed that the results of this meta-analysis were stable. undamaged brain stimulation (MD = 5.87, 95% CI = 2.40; 9.35). The corresponding results are presented in Figure 4B. In addition, a sensitivity analysis showed that the results of this meta-analysis were stable.  Figure 5B. In addition, sensitivity analysis showed that the results of this metaanalysis were credible.

The Meta-Analysis Results for the Functional Oral Intake Scale
Four trials employed the functional oral intake scale [22,23,25,30]. The I 2 statistic = 31% for this group; thus, a fixed-effects model was used to perform the meta-analysis. This meta-analysis showed a significant difference when tDCS was compared with no tDCS (MD = 0.64, 95% CI = 0.52; 0.77) and the corresponding results are presented in Figure 5A. A subgroup analysis result showed that stimulation of the bilateral brain had a positive effect (MD = 0.86, 95% CI = 0.26; 1.46). Stimulation of the undamaged brain had a moderate positive effect (MD = 0.48, 95% CI = 0.21; 0.75). The corresponding results are shown in Figure 5B. In addition, sensitivity analysis showed that the results of this meta-analysis were credible.

The Meta-Analysis Results of the Functional Dysphagia Scale
Three studies used the functional dysphagia scale [23,26,31]. Since the I 2 statistic = 0%, a fixed model was used to perform the meta-analysis. The results revealed that when tDCS was compared with no tDCS, there was a significant difference (MD = −8.15, 95% CI = -13.03; -3.27) and the corresponding results are shown in Figure 6. In addition, sensitivity analysis showed that the trial by Wang (2020) was a major source of heterogeneity. After removing this study, the MD for the functional dysphagia scale was −6.30 [95% CI: −12.74; 0.14, p = 0.0553].

The Meta-Analysis Results of the Functional Dysphagia Scale
Three studies used the functional dysphagia scale [23,26,31]. Since the I 2 statistic 0%, a fixed model was used to perform the meta-analysis. The results revealed that when tDCS was compared with no tDCS, there was a significant difference (MD = −8.15, 95% C = -13.03; -3.27) and the corresponding results are shown in Figure 6. In addition, sensitiv ity analysis showed that the trial by Wang (2020) was a major source of heterogeneity After removing this study, the MD for the functional dysphagia scale was −6.30 [95% CI −12.74; 0.14, p = 0.0553]. Two trials involved Kubota's water-drinking test [24,34]. Since the I 2 statistic = 95%

The Meta-Analysis Results of the Functional Dysphagia Scale
Three studies used the functional dysphagia scale [23,26,31]. Since the I 2 statistic = 0%, a fixed model was used to perform the meta-analysis. The results revealed that when tDCS was compared with no tDCS, there was a significant difference (MD = −8.15, 95% CI = -13.03; -3.27) and the corresponding results are shown in Figure 6. In addition, sensitivity analysis showed that the trial by Wang (2020) was a major source of heterogeneity. After removing this study, the MD for the functional dysphagia scale was −6.30 [95% CI: −12.74; 0.14, p = 0.0553].

The Meta-Analysis Results for Kubota's Water-Drinking Test
Two trials involved Kubota's water-drinking test [24,34]. Since the I 2 statistic = 95%, a random-effects model was used to perform a meta-analysis. When tDCS was compared with no tDCS, there was a significant difference (MD = 0.93, 95% CI = 0.25; 1.61) and the

The Meta-Analysis Results for Kubota's Water-Drinking Test
Two trials involved Kubota's water-drinking test [24,34]. Since the I 2 statistic = 95%, a random-effects model was used to perform a meta-analysis. When tDCS was compared with no tDCS, there was a significant difference (MD = 0.93, 95% CI = 0.25; 1.61) and the corresponding results are shown in Figure 7. In addition, sensitivity analysis showed that the results of this meta-analysis were credible.
corresponding results are shown in Figure 7. In addition, sensitivity analysis showed that the results of this meta-analysis were credible.

The Safety of tDCS
Six studies specified that no skin redness, skin breaks, epilepsy, seizures, headaches, visual disturbances, skin irritation, visual disturbances, or serious adverse events (severe or medically significant but not immediately life-threatening events, include the requirement for inpatient hospitalization or prolongation of hospitalization) occurred [22,26,[29][30][31][32]. The one trial that used the highest intensity stimulation (2 mA) claimed no adverse effects occurred [29]. The other trials did not mention any adverse events.

Publication Bias
Publication bias is a potential concern when interpreting the meta-analysis results. In this study, funnel plots were used to assess publication bias. A publication bias is indicated by an asymmetrical funnel around the pooled effect size. The selected studies did not lie symmetrically around the pooled effect size, as shown in Figures 8-12.

The Safety of tDCS
Six studies specified that no skin redness, skin breaks, epilepsy, seizures, headaches, visual disturbances, skin irritation, visual disturbances, or serious adverse events (severe or medically significant but not immediately life-threatening events, include the requirement for inpatient hospitalization or prolongation of hospitalization) occurred [22,26,[29][30][31][32]. The one trial that used the highest intensity stimulation (2 mA) claimed no adverse effects occurred [29]. The other trials did not mention any adverse events.

Publication Bias
Publication bias is a potential concern when interpreting the meta-analysis results. In this study, funnel plots were used to assess publication bias. A publication bias is indicated by an asymmetrical funnel around the pooled effect size. The selected studies did not lie symmetrically around the pooled effect size, as shown in

The Safety of tDCS
Six studies specified that no skin redness, skin breaks, epilepsy, seizures, headaches, visual disturbances, skin irritation, visual disturbances, or serious adverse events (severe or medically significant but not immediately life-threatening events, include the requirement for inpatient hospitalization or prolongation of hospitalization) occurred [22,26,[29][30][31][32]. The one trial that used the highest intensity stimulation (2 mA) claimed no adverse effects occurred [29]. The other trials did not mention any adverse events.

Publication Bias
Publication bias is a potential concern when interpreting the meta-analysis results. In this study, funnel plots were used to assess publication bias. A publication bias is indicated by an asymmetrical funnel around the pooled effect size. The selected studies did not lie symmetrically around the pooled effect size, as shown in Figures 8-12.

Discussion
Overall, our analysis based on primary outcome measures demonstrated that anodal

Discussion
Overall, our analysis based on primary outcome measures demonstrated that anodal tDCS has a beneficial effect on post-stroke dysphagia, and this result was consistent with previously published studies [16,17]. Moreover, a high intensity, bilateral stimulation tDCS protocol may have a better effect. In contrast to previous studies, our study contains

Discussion
Overall, our analysis based on primary outcome measures demonstrated that anodal tDCS has a beneficial effect on post-stroke dysphagia, and this result was consistent with previously published studies [16,17]. Moreover, a high intensity, bilateral stimulation tDCS protocol may have a better effect. In contrast to previous studies, our study contains all published trials up to 31 December 2021, except for those for which we could not obtain critical and essential outcomes, or the use of electroacupuncture in the control group did not meet our inclusion criteria [35,36]. Many kinds of swallowing function rating scales are used in clinics, such as the dysphagia outcome and severity scale, modified Mann assessment of swallowing ability, functional oral intake scale, functional dysphagia scale, and Kubota's water-drinking test. In this study, we used the dysphagia outcome and severity scale as the primary outcome indicator because it has high reliability [37]. To some extent, our study is valuable because we have included new studies, and our results have updated the stimulation protocol for tDCS. Our most critical finding is that we have used different swallow-related scales to evaluate the effect size of tDCS. In addition, we also reviewed the adverse effects of tDCS on post-stroke dysphagia.
It is currently believed that different polarity, current, or target brain regions for tDCS management would contribute to a wide variety of effects. It has been proposed that tDCS induces neuroplastic changes in motor cortical excitability, i.e., anodal tDCS induces sustained elevations in neural cell membrane potentials, and cathodal tDCS induces sustained decreases in neural cell membrane potentials [38]. In addition, the effect of tDCS may vary according to target brain regions, i.e., the same anodal stimulation may depolarize or hyperpolarize depending on whether the target is in the gyri or sulci, which may explain the large inter-individual variability in tDCS responses [39]. However, recently a new hypothesis has addressed the different effects of polarity, current, and target brain regions on tDCS management that is referred to as the neural noise hypothesis [40,41]. To be specific, the after-effect of tDCS might depend on the overall glutamatergic, GABAergic, dopaminergic, and serotoninergic synaptic activity. Therefore, analysis of the intervention plan of tDCS on post-stroke dysphagia is quite necessary. Thus, we analyzed the effect of stimulation site, stimulus intensity, and other aspects of tDCS on post-stroke dysphagia. The specific details are listed below.

Effect of Stimulation Site of tDCS on Post-Stroke Dysphagia
Subgroup analyses of the modified Mann assessment of swallowing ability and the functional oral intake scale demonstrated that anodal tDCS of the damaged hemisphere and bilateral hemispheres could significantly affect deglutition function in stroke patients. When tDCS is used in different brain areas, it can give rise to various manifestations, for instance, changes in brain networks, cognitive performance, and brain metabolite and neurotransmitter levels [42][43][44]. Since swallowing has bilateral hemispheric representation, the reorganization of the damaged cerebral hemisphere may also play an essential role in recovering deglutition function after stroke [45,46]. For the meta-analysis results of the modified Mann assessment of swallowing ability, the weighted effect size for the bilateral hemisphere was large at 6.19 compared to a medium effect size of 5.87 for the undamaged hemisphere. For the meta-analysis results of the functional oral intake scale, the weighted effect size for the bilateral hemisphere was large at 0.86 compared to the medium effect size of 0.48 for the undamaged hemisphere. These results suggests that anodal tDCS of the bilateral hemisphere is superior to the undamaged hemispheres for improving deglutition function after stroke. Our results are consistent with previous studies showing that the application of tDCS to the bilateral hemisphere may have some inherent advantages over applying it to the undamaged hemisphere. Bilateral tDCS can affect neuronal activity and connectivity within and across the sensorimotor cortical network in the brain [47]. Of course, this result requires further RCT confirmation.

Effect of Intense Stimulation of tDCS on Post-Stroke Dysphagia
The subgroup analysis of the dysphagia outcome and severity scale demonstrated that both low and high-intensity stimulation with anodal tDCS can significantly affect deglutition function in stroke patients. Notably, high-intensity stimulation with anodal tDCS has more advantages than low-intensity stimulation for improving deglutition function after stroke. It is generally known that tDCS is a non-invasive technique that uses a constant, low-intensity direct current (1~2 mA) to regulate neuronal activity in the cerebral cortex. Previous studies had identified that high-intensity (2 mA) stimulation with tDCS had a greater effect on neural plasticity than low-intensity (1 mA) stimulation [48,49]. Here, we divided stimulation with tDCS into high intensity (1.6-2 mA) or low intensity (1-1.5 mA) according to the included studies' characteristics. The results showed that high-intensity stimulation has a better effect size than low-intensity stimulation. As described in previous studies [50], high-intensity stimulation resulted in a significant increase of motor-evoked potentials amplitudes, whereas low-intensity stimulation is also associated with less variability in corticospinal excitability. Moreover, higher cortical excitability is associated with better swallowing function recovery [51].

Duration Stimulation of tDCS
Apart from the intensity of stimulation by tDCS, the duration of stimulation is also an element that can have an impact on the efficacy of tDCS [52]. It has been shown in human studies that tDCS duration varied from 3 to 40 min [53]. In our research, we found that the longest duration of stimulation was 40 min. However, too few durations exist for tDCS for post-stroke dysphagia and it was hard to apply subgroup analyses. Therefore, we were not able to investigate the optimal duration of stimulation.

Treatment Period of tDCS
In this study, the treatment period for tDCS differed between studies. In a previous study, multiple stimulation with tDCS per week may produce a cumulative effect on brain activity and increase its impact on behavioral outcomes [54]. It is generally thought that the short-term and long-term effects of tDCS are different, one of which is resting membrane potential depolarization through non-synaptic mechanisms [55], and the other is N-methyl-D-aspartate-dependent mechanisms [56].

Adverse Effects of tDCS on Post-Stroke Dysphagia
We should recognize that for any stimulation protocol there exists a certain degree of risk that could cause problems in particular cases. Many questions remain open until extensive research or clinical experience is gained. In general, low intensity (1-2 mA) tDCS is considered safe [18]. However, this evidence was collected mainly from healthy subjects and neurological and psychiatric patients. In our research, some studies expressly affirmed that there were no adverse effects reported for post-stroke dysphagia. It should be noted explicitly that one study that used the highest intensity (2 mA) for tDCS declared no adverse events occurred [29]. However, a large sample study of tDCS of healthy subjects and other diseases has reported some negative effects, such as pain, fatigue, itching, etc. [57]. Thus, small sample sizes may explain why the studies included in this review did not report adverse effects.

Limitation
Firstly, this study's data on adverse events was relatively small. As a result, we could not create a quantitative analysis based on the available data; therefore, we could only conduct a narrative review for this study. Secondly, some aspects of the stimulation protocol were different, such as the duration of stimulation and the course of stimulation. Therefore, meta-regression may be needed to adjust these variables. However, the small sample size for those studies limited our ability to do so.

Conclusions
The application of tDCS can promote the recovery of deglutition function in patients with dysphagia after stroke, and bilateral stimulation and high-intensity stimulation may have better effects. However, the safety evidence of tDCS for post-stroke dysphagia is insufficient. In addition, all studies are single-center and lack a unified evaluation scale. Therefore, future research should take steps in this direction to solve these deficiencies.

Conflicts of Interest:
The authors declare no conflict of interest.