Selected articles

A total of 530 studies were identified. After excluding 169 duplicates, the research team reviewed the titles and abstracts of 361 non-duplicate articles. Of these, 343 studies were excluded because they did not meet the criteria (e.g., incorrect population or unrelated to game-based interventions). The remaining 18 articles underwent full-text review, with 2 being study protocols. After excluding these 2 studies, 16 articles, along with 7 additional studies identified through citation tracking, remained for full-text screening. From these 23 articles, 15 were excluded owing to issues with publication type, study design, or intervention type not meeting the inclusion criteria. Ultimately, eight studies were included in this review.

Risk of bias of individuals studies

Figure 3 illustrates the risk of bias for each study. The quality of the studies was assessed as low or high (n = 4 and n = 4, respectively). Low-quality ratings were primarily owing to issues with outcome measurement and the handling of missing data. All studies used a randomization process. Regarding deviations from intended interventions, one study exhibited a high risk of bias owing to unclear reasons for a high dropout rate [29]. In terms of handling missing data, one study had a high risk of bias, as the missing data seemed dependent on the true value of the outcome [30]. Another study showed some concern regarding missing data, but the proportion and reasons for missing data were balanced across groups [29]. For outcome measurement bias, two studies were rated high owing to the absence of validated measurements [31, 32].

Fig. 3

Quality appraisal and summary of included studies

Systematic review of game-based intervention effectsGeneral characteristics of reviewed papers

All eight studies were RCTs. Seven employed a two-arm trial design [29,30,31,32,33,34,35], while one used a three-arm design [36]. These studies were conducted between 2015 and 2024 in seven countries: Belgium, Denmark, Iran, the Netherlands, South Korea, Spain, and the United States. A total of 801 school-aged children and adolescents diagnosed with ADHD participated, with sample sizes ranging from 29 to 348. The age range of the participants in the studies was between 6 and 13 years, and the number of male participants was approximately 4 to 6 times greater than that of female participants in each study. Most studies were conducted in the participants’ homes [29, 31,32,33,34,35,36], with only one conducted in a school setting [30]. The detailed characteristics of the reviewed studies are summarized in Table 1.

Table 1 General characteristics of reviewed studiesObjectives of included studies

Among the studies included in this review, the majority primarily aimed to examine the cognitive effects of game-based interventions for ADHD. Specifically, 62.5% of the studies aimed to improve cognitive processes [30, 32, 34,35,36], 12.5% focused on enhancing daily life functioning [31], and 25% adopted a multi-domain approach [29, 33], targeting both cognitive processes and behavioral symptoms. The most commonly studied cognitive subdomains were response inhibition (n = 5) and working memory (n = 5), followed by cognitive flexibility (n = 3), sustained attention (n = 2), and cognitive control (n = 2). Additional subdomains included information processing speed, multiple simultaneous attention, category formation, and pattern recognition. While none of the studies specifically targeted the psychological effects of game-based interventions for ADHD, two studies assessed socio-psychological outcomes such as quality of life and self-efficacy.

Methodological characteristics and interventions

None of the studies specified a randomization method. Four studies implemented double blinding for the experimental and control conditions [30, 34,35,36], two used single blinding [29, 33], and the remaining two did not use any form of blinding [31, 32]. Most studies (75%) delivered interventions via computer [29,30,31, 33, 34, 36], while two studies (25%) used tablets as the platform [32, 35]. All game-based interventions utilized serious games with mission-guided tasks aimed at improving cognitive function. Participants were required to maintain attention, select targets, retain task-relevant information, and inhibit responses to stimuli. The number of intervention sessions ranged from 20 to 100, conducted over a period of 4 weeks to 3 months. Five studies reported session durations ranging from 15 to 45 min. Over half (62.5%) evaluated participants only at the end of the interventions [29, 30, 32, 33, 35], while 37.5% included follow-up assessments [31, 34, 36]. Among the follow-up studies, two assessed participants at 3 months to evaluate long-term effects, and one study conducted follow-up at 10 weeks. Table 2 summarizes the components of the interventions in each study in detail.

Table 2 Interventions, measurements and key outcomes in reviewed studiesGaming elements and strategies

Supplementary File 1 outlines the gaming elements and strategies used in the reviewed studies. Five studies (62.5%) incorporated rewards, such as points or badges, to motivate participation [31, 33,34,35,36]. Two distinct intervention approaches were identified regarding difficulty level: (a) two studies tailored the workload to individual performance [29, 35] and (b) three studies progressively increased the difficulty to test participants’ adaptive abilities under changing conditions [33, 34, 36]. Medina [29] used an artificial intelligence–driven algorithm to adjust difficulty levels. Three studies provided immersive environments within the games that allowed exploration and meaningful interactions [31, 34, 36]. In two studies by Dovis [34, 36], the main character “Brian” interacted with other game characters, while Kim [31] enabled players to help or communicate with one another via in-game messaging. Using these strategies, Kim [31] uniquely assessed the cooperation subscale as a component of social skills.

Main outcomes

Cognitive functions, including executive function, working memory, inhibition, visuospatial short-term memory, attention, short-term memory, and long-term memory, improved significantly after the intervention. Regarding ADHD behavioral outcomes, findings were mixed: Kim [32] found a significant improvement in behavioral symptoms, while Bikic [33] observed no significant changes. Game-based interventions significantly improved time management and responsibility but did not impact social skills such as cooperation and assertiveness. No studies explored the affective effects of game-based interventions for ADHD.

Meta-analysis results of the game-based intervention effects

Gioia defined executive function as a collection of interconnected yet distinct abilities that regulate cognitive, behavioral, and emotional processes [37]. Based on Gioia’s confirmatory factor analysis of the BRIEF, the following latent factors were identified: (a) metacognition, encompassing initiation, working memory, planning/organization, material organization, and task monitoring; (b) emotional regulation, including shifting and emotional control; and (c) behavioral regulation, which involves self-monitoring and inhibition [37]. The measurements were grouped into cognitive, behavioral, and affective domains (Table 3). A meta-analysis integrating findings from prior studies, using the ELT framework, further examined the effects of game-based interventions across these domains. Figures 4, 5 and 6 present the results of the subgroup meta-analysis.

Table 3 Primary outcomes included in meta-analysisFig. 4

Cognitive results of meta-analysis

Fig. 5

Behavior results of meta-analysis

Fig. 6

Affective results of meta-analysis

Effects on cognitive aspect

The meta-analysis of game-based interventions on cognitive aspects included six subdomains: working memory, initiation, planning/organization, and material organization, measured using the BRIEF, as well as the CBTT Forward and Backward. Compared with the control group, game-based interventions were not effective in improving working memory (MD = − 0.83; 95% CI = [− 2.45, 0.78]), initiation (MD = 1.05; 95% CI = [− 2.14, 4.23]), planning/organization (MD = − 0.53; 95% CI = [− 3.33, 2.26]), or organizing materials (MD = − 0.87; 95% CI = [− 4.24, 2.50]). However, CBTT Forward (MD = 1.67; 95% CI = [0.66, 2.67]) and CBTT Backward (MD = 1.17; 95% CI = [0.16, 2.17]) showed significant improvements after the intervention, indicating that the interventions were effective in enhancing visuospatial short-term memory and visuospatial working memory. Figure 4 presents the results of the meta-analysis on cognitive outcomes.

Effects on behavioral aspects

The meta-analysis of game-based interventions on behavioral aspects included two subdomains: inhibition and monitoring, measured using the BRIEF. Compared with the control group, game-based interventions were not effective in improving inhibition (MD = 0.27; 95% CI = [− 1.44, 1.98]) or monitoring (MD = 1.78; 95% CI = [− 0.92, 4.48]). Figure 5 presents the results of the meta-analysis on behavioral outcomes.

Effects on affective aspect

The meta-analysis of game-based interventions on affective aspects included two subdomains: shifting and emotional control, as measured using the BRIEF. Compared with the control group, game-based interventions were not effective in improving shifting (MD = 0.42; 95% CI = [− 2.84, 3.68]) or emotional control (MD = 0.37; 95% CI = [− 3.16, 3.89]). Figure 6 presents the results of the meta-analysis on affective outcomes.