This study was performed following the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) for protocols (Moher et al., 2009), and the registration number in PROSPERO is CRD42024556058.

The inclusion criteria of this review were as follows: (1) participants must be healthy adults without any psychiatric disorders or musculoskeletal injury; (2) the intervention consisted of at least one session of a-tDCS or/and c-tDCS, and the control group had to be sham-tDCS or no-tDCS; (3) unilateral anodal stimulation, single- or double-blind, randomized experimental design; (4) tDCS overed the prefrontal cortex of the brain; (5) implementation of behavioral tasks, such as SST or/and GNG; (6) measurement of “offline” behavioral changes; (7) English full-text studies published in peer-reviewed journals; and (8) outcome measures should include pre and post stimulation.

Web of Science, Google Scholar, PubMed, Elsevier, Scopus, and Science Direct were searched for appropriate studies published before March 2024. The key words included the following terms: “tDCS” OR “transcranial direct current stimulation” AND “Stop-signal task” OR “Go/no-go task” OR “response inhibition” OR “inhibitory control.” Further relevant studies were manually added by screening the reference sections of retrieved studies, conference papers, and previous reviews.

For each included study, the following data were extracted: (1) study design; (2) sample size and characteristics (gender, age, stimulation interval time); (3) intervention characteristics (duration, current density, electrode size and location); (4) number of sessions; (5) outcome measures (SSRT in the SST, CE rates in GNG); and (6) main results.

The Physiotherapy Evidence Database (PEDro scale) was utilized to assess study quality because of its proven reliability and validity in evaluating randomized controlled trials and ensure a rigorous analysis of evidence in our systematic review and meta-analysis. Two authors (G.Y. and X.P.) independently rated each included study to assess the quality and risk of bias by using the checklist of the PEDro scale (de Morton, 2009); in case the PEDro scale of a certain study was inconsistent, the third experienced author was consulted to reach a consensus. The original scale contains 11 items, which can be used to determine the external (the first item) and internal (the subsequent criteria) validity of the included studies. The results of the PEDro scale evaluation are shown in Table 1. In the present study, we ignored the first question in the study, because it does not match the present purpose. The PEDro score (0–10) of each item was evaluated as poor (< 4), fair (≥ 4 and ≤ 5), good (≥ 6 and ≤ 8), excellent (≥ 9 and ≤ 10) (de Morton, 2009). The total score of each included study was collected for subsequent statistical analysis.

A separate meta-analysis was conducted considering a-tDCS or c-tDCS over the PFC and different tasks for investigating response inhibition by using Review Manager 5.4.1 (Copenhagen: The Nordic Cochrane Centre). Given that some studies used a multi-period stimulation protocol, we decided to extract only the data after the first stimulation. For all included studies requiring quantitative analyses, we entered the three values of mean, SD, and total to calculate the mean difference (MD) and 95% confidence interval (95% CI), which were weighted by inverse variance method using a random-effects model. Study heterogeneity was appraised through the Chi-squared test. The magnitude of the heterogeneity was also quantified using the I² statistic, which was estimated as low (< 30%), moderate (≥ 30% and < 50%), substantial (≥ 50% and < 75%), and considerable (≥ 75% and ≤ 100%) heterogeneity (Higgins et al., 2024). In the case of uncertainty in the information and results provided by the authors and when the data could not be obtained from the figures, tables, and results section, a message would be sent to the first author or corresponding author(s) through e-mail or Research gate to request the mean ± SD of the expected result measurement. If the standard error (SE) was reported instead of SD in the studies, we would use a formula ( S D = S E n , where n is the number of participants in each group) to estimate it (Higgins et al., 2024).

The electronic search yielded a total of 3,182 studies, of which 24 were included in the review. As shown in Figure 1, the selection process is visualized in the PRISMA flow diagram. After the removal of duplicates, 2,430 studies were manually reviewed. A total of 2,267 studies were excluded after assessing titles and abstracts. The full-text articles of the remaining 163 studies were screened based on the eligibility criteria. The 24 included studies comprised a total of 1,228 participants, which were used for qualitative synthesis. Fourteen studies (58.33%) enrolled 714 participants (58.14%) in quantitative synthesis. In the following sections, the effect of a-tDCS and c-tDCS on response inhibition compared with control group would be discussed in each task. The effect of c-tDCS on response inhibition was not used in the meta-analysis because only one eligible study was identified. In addition, the PEDro score of all included studies was between 7 and 9 range, indicating a quality above good level (Table 1).

Four included studies (16.67%) had a total of 3 (6.67%) (Friehs and Frings, 2019), 5 (10%) (Friehs et al., 2021), 5 (12.5%) (Campanella et al., 2018), and 11 (14.10%) (Schroeder et al., 2022) dropouts. The actual sample sizes per study ranged from 11 to 124 participants (51.17 ± 32.71) aged from 18 to 35 years. The study included 484 males and 620 females, only one study lacked report on the gender of the participants (Bell et al., 2020). Three studies (10.53%) only recruited male participants (Campanella et al., 2017; Campanella et al., 2018; Dai et al., 2022), and one study only recruited female participants (Ghayebzadeh et al., 2023).

As shown in Table 2, a randomized design was performed, 16 (66.67%) were parallel and eight (33.33%) were crossover in all included studies. Twenty-three studies (95.83%) had a sham group as a comparator, and only one study (4.17%) had a no stimulation group as a comparator. The current intensity ranged from 0.5 mA to 2 mA, with a stimulation duration of 20.13 ± 5.82 min (ranging from 10–30 min). In terms of electrode placement, the right PFC was selected as the target area for a-tDCS in all but four studies (Fehring et al., 2019; Mansouri et al., 2016; Dai et al., 2022; Schroeder et al., 2022), which selected the left PFC as the stimulation area. Most c-tDCS was placed over the contralateral supraorbital area.

The pooled analysis showed significant changes in response inhibition after single-session a-tDCS compared with the control group in SST and GNG tasks. Figure 2 summarizes the pooled data (size of SSRT) for SST extracted from nine studies on healthy populations. A significant effect in favor of a-tDCS (MD = −17.03 ms; 95%CI = −24.62, −9.43) was found from the random effects model (p < 0.0001), with a moderate heterogeneity (Chi² = 24.83, p = 0.02 and I² = 48%).

As shown in Figure 3, only four studies that analyzed the effect of a-tDCS on response inhibition were found in the GNG task. The pooled MD using the random effects model after a-tDCS was −1.42 (95%CI: −2.71, −0.14, p = 0.03), which indicated that the effects were significant with a low heterogeneity (Chi² = 0.22, p = 0.97 and I² = 0%).

The application of single-session c-tDCS over the PFC led to a significant effect at post-intervention compared with control group (Figure 4). The MD using the random effects model was −15.19 (95%CI: −19.82, −10.55, p < 0.00001) at post intervention with a low heterogeneity (Chi² = 5.18, p = 0.39 and I² = 4%).

The promising outcomes of a-tDCS on modulation executive functioning tasks have recently attracted attention because of its ergogenic potential. As described previously, the right DLPFC is essential for goal-oriented cognitive control tasks (e.g., SST); related studies reported that a-tDCS over the right DLPFC may effectively improve cognitive inhibition processes (Chen et al., 2021; Friehs and Frings, 2018). A growing amount of evidence-based on fMRI studies has indicated that response inhibition is associated with the continuous activation of the right DLPFC that supports the stop process, especially in SST performance (Hughes et al., 2013; Hughes et al., 2014). Interestingly, a-tDCS over the left DLPFC has been confirmed to have a significant ergogenic effect on performing the cognitive task by comparing within (Fehring et al., 2019) and between groups (Mansouri et al., 2016). These findings proposed that a-tDCS can be used to change the susceptibility of the neurocircuitry to regulate cognitive behavior (Fehring et al., 2019). Additionally, accumulating evidence has shown that the right IFC plays a key role in the SST, and a-tDCS over this region can modulate brain activity and enhance functional connectivity to facilitate response inhibition (Cunillera et al., 2014; Jacobson et al., 2011; Sandrini et al., 2020). Functional neuroimaging studies also demonstrated that the right IFC has a large coincidence with a clear overlapped network for inhibitory processes (Chikazoe et al., 2009). Importantly, a-tDCS over the right IFC combined with cognitive training had a more substantial effect on response inhibition (Ditye et al., 2012). The combination of tDCS and cognitive training may emerge as a new method for enhancing cognitive function. However, in disagreement with the present findings, evidence of a previous meta-analysis showed a null effect on inhibitory control while performing a-tDCS over the right DLPFC (Schroeder et al., 2020). This discrepancy may arise from the fact that the previous study included not only healthy participants but also clinical populations. Additionally, the study found that electrode placement, particularly the position of return electrodes, exhibits distance-dependent effects that influence cortical excitability induction. A study has pointed out that the reverse effect of a-tDCS of PFC was mainly because of the differences in neurochemistry and personality for participants (Bell et al., 2020). On a practical and behavioral level, a study documented that the right IFC a-tDCS had a significantly better effect on response inhibition in a single design compared with the right DLPFC; thus, the right IFC is the most reliable brain stimulation target in SST (Stramaccia et al., 2015). The present meta-analysis demonstrated that in the response domain (e.g., SST), the effects of a-tDCS over the PFC studies were significantly larger than sham/no-tDCS groups with moderate heterogeneity.

As described previously, the effects of unilateral a-tDCS of the right IFC on proactive control were significantly increased compared with sham stimulation by testing the GNG task (Cunillera et al., 2014; Campanella et al., 2017; Leite et al., 2018), which may be due to the fact that the inhibition of the promotion response caused by tDCS was related to the decrease in theta band activity (Jacobson et al., 2012; Campanella et al., 2018). However, the anodal electrode was placed on the right IFC combining with left cathodal IFC tDCS, and its ergogenic effects disappeared. A previous study (Leite et al., 2018) speculated that the left IFC was also a particularly critical region in inhibitory control, and the cathodal electrode would weaken the activity of this region and affect inhibition response. The meta-analysis revealed that the application of a-tDCS over the right IFC for clinical populations (e.g., ADHD, major depression, and obesity) can significantly improve proactive inhibition.

To our knowledge, PFC plays an important role in the cognitive control of behavior. The main contribution of the present study is to show the ergogenic response inhibition effects of PFC a-tDCS on reactive and proactive inhibitory processes by means of the SST and GNG tasks, respectively.

Based on the neurophysiological evidence, cathodal offline tDCS can reduce the glutaminergic neuronal activity and the postsynaptic depolarization is weakened so that the cortex produces a long-term inhibitory effect (Stagg and Nitsche, 2011; Cooke and Bliss, 2006). Compared with a-tDCS in response inhibition, a previous study reported that the effect of c-tDCS was smaller in size (Friehs and Frings, 2018). A study by Friehs and Frings (Friehs and Frings, 2019) further showed that c-tDCS over the right DLPFC can diminish the inhibitory effects. By contrast, another study proved that the a-tDCS and c-tDCS of the right DLPFC had a decrease in SSRT for healthy individuals by testing the SST, which suggested that both were effective for response inhibition (Chen et al., 2021). The former pointed out that the inconsistent results could be due to the difference in study procedures (e.g., electrode size and placement, current density and time) and methodology (pre-post or group-between design) (Friehs and Frings, 2019). The possibility of the latter being able to produce an ergogenic result may be the use of a modified SST. Additionally, the latter believed that the significantly decreased SSRT after c-tDCS was due to its promotion of supraorbital activity and indirect adjustment of related frontal cortex through anatomical connections, such as DLPFC and ACC (Chen et al., 2021). Collectively, studies found that compared with sham stimulation, unilateral c-tDCS over the right IFC can reduce SSRT but did not generate a significant effect in SST (Jacobson et al., 2011; Stramaccia et al., 2015). In the included study, only one determined the positive effect of c-tDCS of the right PFC on inhibition response by testing the GNG task (Weidacker et al., 2016). This study concluded that c-tDCS may contribute to regulating the shift of excitatory glutamate to appropriate levels (Weidacker et al., 2016). According to increased shunting over the scalp, the response inhibition network is mainly contained in the right hemisphere (Menon, 2011; Bari and Robbins, 2013). Combined with the present meta-analysis findings, c-tDCS applied over the right PFC can effectively inhibit executive functions with a low heterogeneity under the premise of increasing the sample size. The findings suggest that c-tDCS over the right PFC may be beneficial for clinical populations with inhibitory control deficits. By reducing glutamatergic activity, c-tDCS could enhance response inhibition and could be a potential therapeutic approach for conditions characterized by impaired cognitive control.

Several limitations of this study should be noted. First of all, the participants included in the studies were all healthy populations. Second, the meta-analysis did not perform subgroup analysis in specific regions of PFC. Third, this study did not make a detailed comparative discussion and analysis of stimulation parameters (e.g., current density, duration) and potential moderators (e.g., age, gender). Future studies should investigate the optimal interval time between experimental and control conditions to determine its effect on modulation effectiveness. Finally, the extracted output data were all unilateral offline tDCS after a single session. Future works should use bilateral online intervention to explore the long-term effects of tDCS.