Background
Attention deficit/hyperactivity disorder (ADHD), which commonly presents with behavioral symptoms of inattention and/or hyperactivity/impulsivity inappropriate for a child’s developmental age [
1], is one of the most common neurodevelopmental disorders with a worldwide prevalence of around 5% [
2]. While combined treatment with both pharmacological and behavioral interventions is commonly recommended for children and adolescents with this condition [
3], challenges still exist due to concerns about side effects and stigma with medication use [
4,
5], as well as a lack of satisfactory therapeutic response to behavioral interventions in a significant portion of individuals diagnosed with ADHD [
6]. Moreover, because the symptoms of inattention could persist into adulthood in about 40–60% of children diagnosed with the condition [
7], treatment during adulthood poses another challenge to clinicians due to comorbid presentations of mood problems and psychosocial impairments [
8]. Therefore, in addition to conventional pharmacological and psychosocial interventions, therapeutic approaches of complementary alternative medicine (CAM) are also frequently sought by patients with ADHD and their caregivers [
9].
Neuromodulation, that delivers stimulation to targeted regions of the brain [
10,
11], has been used as a therapeutic alternative for a variety of psychiatric and neurodevelopmental disorders [
12]. Two neuromodulatory approaches, namely repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS), have been shown to offer promising outcomes in the treatment of ADHD [
12]. Although a previous meta-analysis attempted to investigate the therapeutic efficacy of both rTMS and tDCS, the treatment benefits of rTMS remains inconclusive due to lack of available information [
12]. Nevertheless, that study and another meta-analytical investigation both demonstrated preliminary evidence in support of the therapeutic potential of tDCS for improving some cognitive functions such as processing speed and inhibitory controls [
12,
13]. However, the positive findings were only limited to subgroups of selected studies (i.e., tDCS focusing only on the left dorsolateral prefrontal cortex, or in children) [
12], and none of them showed the efficacy of tDCS for improving attentional functions [
12,
13]. In addition, the therapeutic efficacy of rTMS, which involves induction of a more focal stimulation than that with tDCS [
14], has not been systematically reviewed for ADHD-related symptoms and cognitive deficits [
12].
The potential therapeutic effects of rTMS are theoretically based on neurostimulation targeting areas of hypofunction in specific brain regions responsible for different cognitive functions [
12]. While the exact mechanism underlying the therapeutic effects of rTMS remains unclear, several studies have reported an association of rTMS treatment with an induction of brain dopamine release [
15] and an increase in synaptic plasticity [
16,
17]. Despite the choice of the prefrontal cortex (PFC) as the target region for rTMS stimulation in previous randomized controlled trials (RCTs) [
18‐
20], taking into consideration its critical role in attentional as well as many other important cognitive functions [
21], their findings regarding the therapeutic effectiveness of rTMS varied with different core symptoms of ADHD [
18‐
20]. For instance, one study showed an efficacy of rTMS only for the symptoms of inattention but not those of hyperactivity/impulsivity [
18]. That study further demonstrated that rTMS was effective for improving inattention when targeting the right PFC but not the left PFC [
18]. The finding was consistent with that of a previous study showing predominant underactivation of brain regions on the right hemisphere rather than those on the left side [
22]. On the other hand, the previous report of an association between the left brain hemisphere (i.e., left caudate head) and impulsivity suppression [
22], may explain the efficacy of tDCS targeting the left dorsolateral PFC for better inhibitory control in previous meta-analyses [
12,
13]. These findings, therefore, highlighted the probability that the therapeutic effects of rTMS may vary when targeting different brain regions [
18].
Informant bias remains an unabated issue when assessing the therapeutic outcomes for children with ADHD [
23]. Indeed, subjective rating of clinical symptoms not only may be more susceptible to informant biases [
24], but also could not provide detailed information regarding specific neuro-cognitive functions [
25]. Therefore, objective evaluation of different cognitive functions with cognitive tasks may be more reliable in the assessment of the therapeutic efficacy of rTMS in patients with ADHD.
The aims of present meta-analysis were, therefore, to investigate the therapeutic efficacy of rTMS focusing on different cognitive performances measured by standardized tasks as well as to identify possible factors (e.g., different locations of stimulation) that may influence its therapeutic outcomes in patients with ADHD.
Discussion
Although treatments using brain stimulation have been reported to improve cognitive functions and attention associated with several psychiatric and neurodevelopmental disorders [
32], evidence supporting the therapeutic effect of rTMS on attentional performance was still limited [
12]. To our best knowledge, our meta-analysis is the first to investigate the therapeutic efficacy of rTMS for improving cognitive functions in patients with ADHD. Our results demonstrated more significant improvements in sustained attention and processing speed but not in memory and executive function among ADHD patients treated with rTMS compared with those in the control groups. On the other hand, we found no significant difference between rTMS focusing particularly on rPFC and that targeting other locations (e.g., both right and left PFC or lPFC only). Nevertheless, the soundness of evidence derived from the current study warrants elucidation from further clinical investigations given the limited number of eligible studies (
n = 5, total sample size = 189).
The result of our primary outcome showed that treatment with rTMS was associated with a more significant improvement in sustained attention than in controls when assessed by mindstreams battery and CPT, which were computerized versions of standardized attentional tests. Given that standardized attentional tests may provide a more objective assessment, our results were less susceptible to informant bias [
24], which has been reported to be an significant issue in studies using questionnaire-based evaluation for patients with ADHD [
33]. Moreover, assessment of cognitive function by using standardized attentional tests may better reflect improvement in neurocognitive functioning (i.e., endophenotype) [
25] that should theoretically precede behavioral improvements (i.e., phenotype) reflected by questionnaire-based rating scales. Therefore, based on standardized attentional tests, our results may be more sensitive to the detection of the therapeutic effects of rTMS on neurocognitive functions compared to questionnaire-based studies. In addition to a minimization of informant bias, low risk of performance bias in our results was also evident as reflected by the consistency of our primary outcome (i.e., improvement in sustained attention) on leave-one-out sensitivity analysis after excluding the only study that did not use sham rTMS [
30]. Overall, our results supported the therapeutic efficacy of rTMS for improving attentional function in patients with ADHD.
Despite unclear mechanisms underlying the therapeutic effects of rTMS on neurocognitive functions, previous studies have demonstrated an association of rTMS treatment with induction of dopamine release in the brain [
15] and enhancement of synaptic neuroplasticity [
16,
17]. It is also well-reported that different regions of the brain are linked to various cognitive functions [
34]; therefore, rTMS targeting specific brain regions to enhance a particular neurocognitive function seems a reasonable therapeutic approach [
32]. Indeed, previous evidence suggested a stronger correlation of cognitive dysfunction with the right hemisphere compared to that on the left in patients with ADHD [
35,
36], highlighting the importance of investigating the possible influence of targeting different brain regions on the therapeutic efficacy of rTMS. Nevertheless, notwithstanding the report of a stronger association between attentional function and the right dorsal prefrontal cortex (rDLPFC) in a prior meta-analysis of brain imaging data from patients with ADHD, the same study also suggested that interference inhibition was more related to the left caudate head [
36]. Taking into account the need for inhibiting irrelevant stimuli to maintain concentration [
37], both left and right hemispheres of the brain may have important albeit different roles to play in sustaining attention. Consistently, targeting the left side of the brain was found to enhance inhibitory control and processing speed in patients with ADHD in previous meta-analyses investigating tDCS, which is another electrical stimulation approach [
12,
13]. However, we could only study the therapeutic efficacy of rTMS targeting rPFC but not those focusing on the left hemisphere due to the limited number of RCTs targeting the left hemisphere (
n = 1) [
31]. While our subgroup analysis showed similar therapeutic efficacies for improving sustained attention between rTMS targeting rPFC and other approaches (e.g., those targeting both right and left PFC), we were unable to study inhibitory controls due to a lack of available outcome data from the eligible studies. Therefore, although our subgroup analysis failed to demonstrate significant differences in therapeutic efficacy of rTMS for improving sustained attention through targeting different brain regions, we could not rule out the possibility that targeting different brain regions may enhance different aspects of attentional functions.
With regard to our secondary outcomes, the results were only representative of the therapeutic efficacy of rTMS targeting rPFC because only studies focusing on rPFC provided analyzable data. Our findings showed that the therapeutic efficacy of rTMS was only significantly better than that of the controls in terms of processing speed, but not in memory and executive functions. However, leave-one-out sensitivity analysis showed significantly superior therapeutic efficacies of rTMS for improving memory and executive function after excluding the study by Bleich-Cohen M et al. [
18]. Nevertheless, the result was derived from only two remaining studies [
20,
30]. Besides, a significant heterogeneity was noted in the result for memory (I
2 = 71%,
p = 0.03). Therefore, the results of our secondary outcomes regarding memory and executive function were not robust and warranted further studies for elucidation. On the other hand, our finding of rTMS-related improvement in processing speed was consistent with that of another meta-analysis on ADHD patients treated with tDCS [
13]. Nevertheless, the positive result on processing speed improvement in that meta-analysis came from tDCS targeting the left PFC instead of the right as reflected in our finding. In addition, several fMRI studies also found that processing speed may be more associated with the left PFC [
38,
39]. On the other hand, some brain-imaging studies have shown that both right and left PFC may play a role in processing speed [
40,
41]; therefore targeting rPFC may still be a therapeutic option for improving processing speed. The limited available data precluded our investigation into the difference in therapeutic efficacy between rTMS targeting the right PFC and that targeting the left PFC. Besides, because rTMS-elicited changes in processing speed may be attributed to improvements in different domains of cognitive functions, further studies are warranted to shed light on these issues.
Despite the design of the current study to minimize informant bias and provide more reliable measures of neurocognitive functions, there were several limitations. First, notwithstanding the moderate ES regarding the improvement in sustained attention related to rTMS treatment without significant heterogeneity or a change in result on sensitivity analysis, the significance of our finding may still be obscured by the limited number of included studies and participants. In addition, the measurement of heterogeneity by using I-squared statistic is prone to bias, especially when the numbers of studies are small. Second, while the majority of our included studies recruited adult participants diagnosed with ADHD, only one targeted children and adolescents [
30]. Therefore, taking into account the difference in stages of brain development between adults and children/adolescents [
42], the results of this meta-analysis may not be generalizable to the younger populations. For instance, one study still found differences in the areas of brain activation between the two populations during tasks for verbal working memory despite their overall similarity [
42]. Third, despite previous evidence showing an association between ADHD and the right PFC [
35,
36], we were unable to directly compare the treatment efficacies between rTMS targeting the right PFC and that targeting the left side. Nevertheless, our subgroup analysis demonstrated quite similar ES between rTMS only targeting rPFC and rTMS targeting other locations (i.e., both left and right PFC or left PFC). Given that inhibitory control, which is quite important for sustaining attention, has been reported to be more related to the left hemisphere [
36], current evidence is still insufficient to provide more precise information about the influence of targeting different brain regions on different subcategories of cognitive functions. Fourth, the limited geographic scope of our included studies (i.e., three from Israel [
18‐
20], one from China [
30], and one from USA) may restrict extrapolation of our findings to populations of different ethnic and cultural backgrounds. Fifth, our results derived from only objective assessments of cognitive functions may not reflect patients’ perception or subjective feelings from important others (i.e. parents or teachers). Nevertheless, results from objective evaluations of cognitive functions may be better indicators of improvement in neurocognitive functions compared with those acquired with subjective tests [
25]. Finally, we were unable to conduct subgroup analyses or meta-regression focusing on other important factors (e.g., age or number of treatment sessions) that could affect the therapeutic outcomes of rTMS due to the limited number of RCTs included in current meta-analysis.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.