Impulsivity is a multifaceted construct that reflects a tendency to act rashly and without forethought, with negative consequences for the individual and others (1, 2). Trait impulsivity refers to personality traits and behaviors encompassing urgency, sensation seeking, lack of perseverance, and disinhibition (3). Behavioral manifestations of impulsivity are multifaceted, including deficits in response inhibition (motor), delay discounting (temporal), and information sampling (reflection or cognitive). Additionally, Impulsivity overlaps significantly with decision making (4).

Several neuronal circuits in the brain have been implicated in impulse control. Motor impulsivity is regulated by a fronto-subcortical circuit involving the right inferior frontal gyrus, the anterior cingulated cortex, the basal ganglia, and the presupplementary motor area (5, 6). The ventromedial prefrontal cortex has been implicated in cognitive impulsivity and decision making under conditions of risk or ambiguity (7, 8). In contrast, temporal impulsivity is regulated by a fronto-limbic circuit including the anterior cingulated cortex, the ventromedial prefrontal cortex, and the nucleus accumbens of the ventral striatum (9). Dysfunction in the thalamo-cortico-striatal neural network has also been implicated in impulsivity (10).

Trait impulsivity is measured using questionnaires such as the Barrett Impulsiveness Scale [BIS-11; (11)] and the Urgency-Premeditation-Perseverance-Sensation Seeking-Positive Urgency (UPPS-P) impulsive behavior scale (3), whereas state impulsivity is measured using behavioral measures such as the Stop Signal Task (12), Delay Discounting Task (13), Go-No-Go tasks (14), and Information Sampling Task (15).

Furthermore, impulsivity is a key feature of many of the externalizing disorders listed in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders [DSM-5; (16)] including attention deficit hyperactivity disorder, antisocial personality disorder, borderline personality disorder, substance use disorder, and pathological gambling. It is also an important consideration in risk assessment tools owing to its association with aggression and suicidality (17). Impulsivity along with craving, and neurocognitive deficits have also been implicated in maintaining addictive behaviors in people with substance use disorder (18, 19). Additionally, impulsivity has been associated with poor treatment adherence in people with cocaine use disorder (20).

A range of interventions have been developed to reduce impulsivity in the context of addiction including computerized cognitive training, cognitive remediation, pharmacological interventions (e.g., modafinil, galantamine), and mindfulness relapse prevention (21, 22) with some promising results. Cognitive control training has been also showed to reduce emotion-related impulsivity in adults (23). Additionally, a range of pharmacological interventions are used in clinical practice to reduce impulsivity, with some evidence supporting the use of fluoxetine, carbamazepine, and topiramate for impulse control disorders (24) and quetiapine for impulsivity in people with borderline personality disorders (25).

Non-invasive brain stimulation techniques (henceforth referred to as NIBS) include transcranial direct current stimulation (tDCS), transcranial random noise stimulation (tRNS), transcranial alternating current stimulation (tACS), transcranial magnetic stimulation (TMS), deep brain stimulation (DBS), and others. While tRNS and tACS are primarily used as investigative tools in research, tDCS and TMS have been used for the treatment of various neuropsychiatric disorders such as depression, obsessive compulsive disorder, substance use disorder, schizophrenia, and Parkinson’s disease (PD) amongst others (26–28). Additionally, data support the use of tDCS to improve substance use-related outcomes by reducing craving and relapse rates (29), highlighting its potential as a non-pharmacological option for substance use disorders (30). Moreover, the utility of NIBS techniques in reducing impulsivity has been demonstrated in systematic reviews in both healthy subjects (31) and people with mental disorder (32), although the results are often inconsistent across studies. While the exact mechanism of action of NIBS is not fully understood, their effects are thought to be related to inducing neuroplasticity changes in the brain (33) and modulating the function of various neurotransmitters (34).

According to Gobet and colleagues (35), “Cognitive training refers to interventions using cognitive tasks or intellectually demanding activities, the goal of which is to enhance general cognitive ability.” Cognitive training encompasses a wide range of activities including performing cognitive tasks, learning music, and playing videogames (35). Central to cognitive training are the concepts of near transfer (i.e., the generalization of acquired skills across two or more related domains) and far transfer (i.e., the generalization of acquired skills across loosely related domains) (35, 36).

The utility of cognitive training in promoting neuroplasticity changes in the brain has been demonstrated in addiction disorders (37). Cognitive training has been shown to have a modest effect on enhancing cognitive functioning in people with mild to moderate PD especially in areas of memory, processing speed, and executive functions (38). Cognitive rehabilitation has been shown to improve cognitive performance and resting functional brain connectivity and reduce functional disability in PD (39). Despite growing interest in cognitive training and a multi-billion-dollar industry, empirical evidence in the field remains inconclusive (35, 36).

There is some evidence to support the use of cognitive training to enhance the effects of tDCS both on trained tasks and non-trained but related tasks (30, 40). Indeed, the neuroplasticity-related effects of tDCS are enhanced when tDCS is administered over an already engaged brain region while engaged in a cognitive training task (41). For example, a recent review reported small but statistically significant effects for combined cognitive training and tDCS on attention, working memory, language and global cognitions in people with neuropsychiatric disorder (42). Although previous reviews showed positive effects on cognitive functioning for combined Transcranial Electrical Stimulation (tES) and cognitive training (43) as well as combined cognitive training and tDCS (42), it is unclear if the putative advantage of combing cognitive training and NIBS will hold true for impulsivity. This is important since the existing evidence suggests that combined cognitive training and tDCS, for example, can have a synergistic positive effect on behavior (44–46).

Given the potential negative consequences of impulsivity (1) and the absence of published reviews that specifically examined the effects of combined NIBS and cognitive training on impulsivity, we aimed to conduct a systematic review and meta-analysis to fill this gap in the literature. Conducting a review of this kind would help not only to enhance knowledge, but also to identify areas for future research with a view to developing adjunctive interventions for impulsivity. Although a range of pharmacological and psychosocial interventions have been developed to target impulsivity in various disorders, these are not without limitations. Take attention deficit hyperactivity disorder as an example. A longitudinal study of ten European countries showed that while medications showed positive effects on attention deficit hyperactivity disorder symptoms, the effects of psychosocial interventions were either insignificant or negative (47). Although more recent studies highlighted the beneficial effects of psychosocial interventions (48) and demonstrated relatively large effect sizes in the short-term in populations with attention deficit hyperactivity disorder, there is still a need to enhance current therapeutic strategies for this patient population (49). Put together, these findings emphasize the need to develop adjunctive interventions, such as combined cognitive training and NIBS, for disorders that are marked by impulsivity.

The review protocol is enrolled in the PROSPERO international prospective register of systematic reviews (CRD42024511576) and reporting follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (50).

The PICO framework was used to determine inclusion criteria. Studies were included in the review if they (i) involved adult participants with or without mental disorder; (ii) involved the use of combined cognitive training and active NIBS (e.g., tDCS, TMS, DBS, tES); (iii) included a control group such as combined cognitive training and sham NIBS; and (iv) used at least one behavioral tool to measure overall impulsivity or its facets and reported the findings of these tools. We included studies involving participants with mental disorder since impulsivity is a core feature of some of the mental disorders classified in the Diagnostic and Statistical Manual of Mental Disorders fifth edition (DSM-5) including attention deficit hyperactivity disorder, borderline personality disorder, and antisocial personality disorder (16). Studies involving children and adolescents, editorials, non-controlled studies, and those in languages other than English were excluded. No other specific exclusion criteria were applied for setting, region, or date of publication.

The initial search covered the period August 1980 until September 2024 and included Scopus, PsychInfo, Medline, and CINHAL electronic databases in addition to dissertations databases (ProQuest Dissertations and Theses Global, Open Access Theses and Dissertations) and Google Scholar. The lists of references of the included studies were hand searched to identify studies not picked up by the initial search. Grey literature was also explored using Google.

We defined cognitive training in accordance with previous studies [e.g., (35, 42, 51)], including cognitive tasks designed to enhance specific domains such as working memory, decision making, attention, executive function, and others. We conceptualized impulsivity and its subdomains (e.g., motor, temporal, and cognitive or reflection impulsivity) in accordance with previous literature [e.g., (4, 52, 53)]. Appropriate behavioral measures were identified to measure motor impulsivity (e.g., Stop Signal Task, Go-No-Go task, and the Stroop Color and Word Test), temporal impulsivity (e.g., Delay Discounting Task), and reflection or cognitive impulsivity (e.g., Information Sampling Task).

The search terms were adapted from previous studies (4; e.g., 42) and included, “Transcranial magnetic stimulation”, “TMS”, “theta burst stimulation”, “TBS”, transcranial Direct Current Stimulation”, “tDCS”, “transcranial Electrical Stimulation”, “tES”, Random Noise Stimulation”. “RNS”, “transcranial Alternate Current Stimulation”, and “tACS”, COMBINED with “cognitive or memory” adjacent to (train*” or remediate* or enhancement* or rehabilitate* or treatment* or therapy*) AND “impulsive*”, “self-regulation”, “inhibitory control”, “impulse control”, “delay discounting”, “response inhibition”, “information sampling”, “stop signal”, “temporal discounting”, “stroop”, “go-no-go”.

Using the Covidence workflow platform (http://www.covidence.org), three reviewers independently screened titles and abstracts to identify potentially eligible articles. Disagreements were resolved by consensus. Full texts of potentially eligible articles were reviewed to identity a final list of studies for inclusion.

Data extraction was conducted using a data collection tool designed for this study which gathered information about study title, publication date, country of origin, setting, study design, participant characteristics (gender, age), interventions, and outcome measures as well as study limitations.

Data were pooled using random-effects meta-analysis in the Statistical Package for Social Studies (SPSS) version 27. Random-effects meta-analysis models assume that treatment effects vary across studies due to differences between the studies in addition to sampling variability. This is appropriate due to the high degree of clinical heterogeneity across the studies included in this review (54).

Evidence suggests that impulsivity is heterogenous, and that motor-, temporal-, and reflection-impulsivity should be considered separately, since the different domains reflect different types of behaviors and thus different underlying processes (55). Since subdomains of impulsivity have distinct neurobiological underpinnings and owing to differences in the mechanism of action of NIBS techniques, we aimed to conduct separate metanalyses by domains of impulsivity and type of NIBS technique to minimize heterogeneity where possible.

The principal effect size was represented as Hodge’s (adjusted) g and 95% confidence intervals (CI) by calculating the mean differences between experimental (combined cognitive training and active NIBS) and control (combined cognitive training and sham NIBS sham) conditions in post-stimulation evaluations divided by the pooled standard deviation (a summary measure known as the standardized mean difference) multiplied by a correction factor for small samples. Effect sizes were considered negative if the active intervention was in the predicted direction, and positive otherwise. If a study used multiple NIBS stimulation sites, each simulation site trial was used as the unit of analysis in meta-analysis. To minimize the risk of heterogeneity, reaction time was used if studies provided multiple outcomes (e.g., reaction time, subscale scores, performance, or accuracy). Additionally, only data relating to the immediate post-stimulation time point were included in the analysis in cases where outcome measures were assessed at different time points. Only one study in this review measured outcomes at more than one post-stimulation time point (56). For studies involving more than one control condition, only data pertaining to combined active NIBS and cognitive training and combined sham NIBS and cognitive training were included in the analysis.

Between-study heterogeneity was assessed using the H² , Q, I² , and tau² statistics (57). The I² statistic reflects the percentage of variation related to heterogeneity rather than chance. The Q statistic represents the weighted sum of squared differences between individual study effects and the pooled effect across studies, while H² represents the relative excess in Q over its degree of freedom (58). In contrast, the tau² statistic is represents the variance of the true effect sizes (59). Heterogeneity values are described in the Forest plot ( Figure 1 ). For this review, I² values of <25% were considered ‘low’; I² values >40% and p values of ≤0.05 were considered ‘moderate,’ and I² values >75% and p values of ≤0.05 were considered ‘high’.

Funnel plots were used to test for the presence of publication bias. Where publication bias was detected, the Trim and Fill procedure was applied to examine study clustering and attempt to correct for this (60) while acknowledging that a symmetrical funnel plot and non-significant tests for funnel plot asymmetry do not necessarily exclude the possibility of publication bias.

Quality assessment was conducted using the NIH National Heart, Lung and Blood Institute’s Quality Assessment Tool for Controlled Intervention Studies (61). Each study was assessed against predefined criteria that covered the domains of power calculation, randomization, allocation concealment, blinding, adherence to interventions, dropout rates, intention-to-treat analysis and others. Each study was assigned an overall quality rating (good, fair or poor) based on the total number of criteria that each study met and the potential impact of not meeting these criteria on the results.

The initial search identified 123 records. After removing duplicates, 102 records remained for screening of which only 4 studies were included in the review. The PRISMA flow diagram provides more information about the rationale for exclusion ( Figure 2 ).

Table 1 summarizes study characteristics. Four randomized controlled studies involving the use of combined cognitive training and tDCS and 127 subjects were included in the study. Of those, two studies examined the efficacy of combined cognitive training and tDCS for substance use disorders such as opioid use disorder (62) and methamphetamine use disorder (56), one for Parkinson’s disease (63), and one for obesity (64). No none of these studies enrolled participants primarily based on impulsivity or used impulsivity as a primary outcome measure. All four studies targeted the dorsolateral prefrontal cortex (DLPFC) for tDCS stimulation. A range of cognitive training tasks were used including the Game of Dice Task, Brain HQ Computerized Cognitive Training, and computerized cognitive addiction therapy. Impulsivity was measured using various behavioral tasks such as the Stop Signal Task, Go/No Go task, Dot Counting Task, and Dot Probe Task. A wide range of other tools were used to measure cognitive functioning, mental health symptoms, global functioning, and quality of life.

Table 2 provides a summary of overall quality ratings. Three studies were rated as ‘good’ (56, 62, 63) and the remaining one (64) as ‘fair.’

Tests of heterogeneity were insignificant (Tau² = 0.00, H² = 1.00, I² = 0.00). the funnel plot showed no evidence of publication bias ( Figure 3 ). However, clinical heterogeneity related to the inclusion of individuals with different neuropsychic disorders and those with obesity was evident across the studies.

Given that facets of impulsivity correlate weakly with each other and have distinct neurobiological underpinnings (4), we elected to separately examine the effects of combined cognitive training and tDCS on different facets of impulsivity. Two studies measured motor impulsivity using the Stop Signal task (62) and a Go/No Go task (63). The statistical results are presented below under the relevant headings. It is worth noting that beyond statistics, we did not observe any specific trends across the different studies in relation to impulsivity outcomes.

Meta-analysis of the overall effects of combined NIBS and cognitive training on motor impulsivity or inhibitory control showed that the effects were insignificant (g= -0.15, 95% CT, -0.50 to 0.20, p=0.40) See Figure 1 .

Since the Flanker Task measures both attention and inhibitory control, we conducted a separate meta-analysis including SST and Go/No Go tasks. The results showed no statistically significant effects for combined NIBS and cognitive training on task performance (g= -0.23, 95% CT, -0.60 to 0.14, p=0.22) ( Figure 4 ).

One study reported pre and post stimulation data on delay discounting (56). The results showed no statistically significant difference in discounting scores between combined active tDCS and cognitive training and combined sham tDCS and cognitive training groups.

None of the studies included in this review examined the effects of combined NIBS and cognitive training on reflection of cognitive impulsivity, highlighting the need for conducting studies of this kind.

A major strength of this study is its reliance on meta-analytic techniques to examine the effects of combined cognitive training and non-invasive brain stimulation on impulsivity and its subdomains. At the same time, the review was limited by the relatively small number of studies included in the review, which mainly involved males with substance use disorders. Notably, studies included in this study mostly focused on motor impulsivity and delay discounting and no studies measured reflection or cognitive impulsivity. Additionally, no study included participants based on impulsivity or used impulsivity as a primary outcome measure. These in conjunction with clinical heterogeneity limit the generalizability of findings to other populations such as those with other forms of major mental disorder and overall conclusions about impulsivity subdomains. The low sample size (n = 127 across four studies) additionally limits the statistical power and interpretation of the meta-analysis. Effect sizes may be under reported, and any biases in the included studies may be amplified. Finally, the search strategy included Scopus, PsychInfo, Medline, and CINHAL electronic databases in accordance with a recent review on combined cognitive training and tDCS (42). This may have missed studies listed in other relevant databases such as EMBASE.

Nascent literature in the field shows that tDCS and TMS, either as standalone interventions or combined with neurorehabilitation therapies, may positively alter neuroplasticity and improve neuropsychological, neuropsychiatric, motor, or somatic symptoms through brain stem activation (75). This mechanism, also known as the hormesis principle of neuroplasticity, has been proposed for explaining the use of NIBS to improve maladaptive brain physiology and behavioral symptoms resulting from acquired brain injury (76). Moreover, a recent imaging study identified reduced complexity in vmPFC as a putative mechanism for impulsivity in terms of risky and impatient economic choices (77). These putative mechanisms merit further investigations in studies involving the use of combined cognitive training and NIBS to modulate impulsivity and its subdomains. Future research should focus on conducting well designed studies involving disorders that are marked by impulsivity and including mechanistic evaluations such as brain imaging techniques, electroencephalography, and neurocognitive measures.