BCI games that are currently used for cognitive assessment and in which populations

Most studies in this review focused on healthy male adults, highlighting an underrepresentation in participant characteristics across age, sex and clinical status. The average age across studies is around 26.5 years, with pediatric samples being limited and fragmented, ranging from 3.5 to 16 years old. In terms of sex distribution, a marked imbalance was evident; most studies skewed heavily toward male participants, with some samples being 100% male. Most participants were healthy individuals (76%), with smaller subsets including individuals with ADHD (18%) and other conditions. Notably, the overrepresentation of ADHD in pediatric samples (57% of children participants) further narrowed the scope of BCI research, reinforcing a focus on attention-related assessments, a pattern likely driven by convenience sampling and the natural fit of attention tasks with EEG paradigms. While neurotypical adults offer a convenient starting point for assessing BCI game usability and performance, they do not represent the full spectrum of potential end-users, especially those who benefit most from nonverbal, adaptive cognitive assessments. This limits the external validity and real-world applicability of BCI game findings, underscoring the need for more diverse recruitment strategies and standardized sex-disaggregated reporting in future studies. ^{Reference McEwan62–Reference Andrade64}

The types of games used across studies varied, but a clear pattern emerged. Nearly half of the games were training games designed to target attention or working memory through repetitive tasks, neurofeedback and adaptive difficulty (e.g., FOCUS, REEFOCUS). ^{Reference Georgiou, Thanos, Kanellos, Doulgerakis and Thomopoulos35,Reference Alchalabi, Elsharnouby, Shirmohammadi and Eddin36} Action/adventure games incorporated fast-paced challenges or exploratory tasks (e.g., FarmerKeeper, BrainCats). ^{Reference Mercado, Espinosa-Curiel, Escobedo and Tentori40} Puzzle games involved logic, memory matching or spatial manipulation (e.g., Scalebridge, text rotation games). ^{Reference Tahai, Wallace, Eckhardt and Pietroszek65,Reference Kong and Song66} Less frequently, studies used simulation games (e.g., AVERT) to recreate real-world tasks or narrative-driven games (e.g., fairy-tale games) to engage users in story-based interactions. ^{Reference Kenny, Veitch and Power67} Most games relied on attention as the primary control command, with players modulating game elements such as speed, movement or task success based on EEG signals. A smaller proportion of studies combined EEG with other modalities, such as motion sensors or eye-tracking, to create more immersive, multimodal gaming experiences.

Cognitive and perceptual skills measured by BCI games, and how were they measured

The adaptability of BCI games stands out as a major strength, offering a flexible, user-centered approach to cognitive assessment. This is particularly relevant for populations with motor or communication challenges, where traditional methods often fall short. However, the narrow focus on specific cognitive domains, predominantly attention, limits the broader applicability of current findings. The overreliance on attention as the primary cognitive skill assessed in BCI games is a critical finding of this review. This narrow focus is directly tied to the participant demographics, particularly the overrepresentation of children with ADHD in pediatric samples. The suitability of EEG-based paradigms (e.g., theta/beta ratio) for detecting attentional variations makes this a theoretically grounded starting point for BCI research; however, it has inadvertently limited the scope of the field. ^{Reference Herrera-Morales, Reyes-López and Tuz-Castellanos68,Reference Bluschke, Broschwitz, Kohl, Roessner and Beste69} Other cognitive domains are equally tractable via EEG: working memory can be indexed by theta oscillations and P300 amplitude, cognitive flexibility by frontal alpha asymmetry and error-related negativity and problem-solving by gamma-band synchronization. ^{Reference Klimesch58,Reference Scharinger, Soutschek, Schubert and Gerjets70–Reference Sheth, Sandkühler and Bhattacharya72} BCI games targeting these domains could be validated against established neuropsychological tools, providing a principled psychometric bridge between neural signals and clinically recognized constructs. A more deliberate effort to expand beyond attention is needed to demonstrate the full potential of BCI games as comprehensive cognitive assessment tools. Future research should prioritize designing and validating BCI games for a broader range of cognitive domains, such as memory, problem-solving and executive function, to broaden the clinical applicability of this technology across diverse populations. In the few studies involving older adults, the focus shifted slightly toward attention and memory, reflecting the relevance of these domains for age-related cognitive decline and conditions like MCI. ^{Reference Saunders and Summers73}

Moreover, only a few studies have assessed skills such as memory, problem-solving or cognitive flexibility, highlighting a need for more comprehensive assessment frameworks. These findings align with previous studies, emphasizing the promise of task-specific BCIs for cognitive enhancement while noting protocol standardization and user adaptation challenges. ^{Reference Simon, Bolton, Kennedy, Soekadar and Ruddy74} Similarly, other studies highlighted the role of gamification in improving engagement and stressed the importance of robust validation against standardized assessments – a recurring theme in this review. ^{Reference Lumsden, Edwards, Lawrence, Coyle and Munafò14,Reference Barends, de Vries and van Vugt75–Reference Gomez, Ruiperez-Valiente and Clemente77}

The methods used to measure these cognitive skills varied as well. Most studies combined game performance metrics (e.g., accuracy, reaction time, errors) with EEG signals, allowing researchers to infer neural correlates of cognitive states while also capturing behavioral outcomes. This combined approach enhances the ecological validity of assessments, as it links brain activity with task performance in real time. ^{Reference Chaytor, Temkin, Machamer and Dikmen78,Reference Ewing, Fairclough and Gilleade79} However, a small number of studies relied solely on EEG data or game performance, which may limit the depth and reliability of cognitive insights, particularly in complex domains such as memory or problem-solving. A few studies incorporated additional sensors, such as eye-tracking and motion sensing technologies, to create multimodal BCI systems. These setups aimed to capture a more holistic profile of user engagement, combining neural, behavioral and physical data streams. While promising, these multimodal approaches also introduce technical challenges, including comfort issues with wearing multiple devices, data synchronization problems and potential signal interference between sensor systems. ^{Reference Casson80}

BCI games for cognitive assessment compared to traditional methods

Compared to traditional cognitive assessments, BCI games are emerging as promising alternatives that combine neurophysiological monitoring with interactive, user-centered design. Most studies reviewed demonstrated moderate-to-strong correlations between BCI game performance and standardized cognitive assessments, supporting their validity for evaluating cognitive functions. These findings suggest that BCI games hold promise as engaging tools for cognitive training, particularly in attention-related tasks, and may offer enhanced engagement compared to traditional assessments. Several studies report that BCI games increase participant motivation, enjoyment and sustained involvement, especially in children with ADHD, suggesting a potential advantage in user engagement. ^{Reference Georgiou, Thanos, Kanellos, Doulgerakis and Thomopoulos35,Reference Park, Kihl, Park, Kim and Chang41,Reference Rajabi, Pakize and Moradi81–Reference Chen, Tang, Zhang and Shin83} Real-time feedback, gamified environments and the ability to tailor difficulty levels can enhance user engagement, contribute to higher classification accuracies and improve the quality of EEG signals by reducing movement artifacts caused by boredom, fatigue or inattention. When users are motivated and engaged, they are more likely to maintain focus and minimize erratic movements, resulting in cleaner and more stable EEG patterns that enhance the reliability of classification. These features align with previous research that emphasizes the role of gamification in improving motivation and data quality in cognitive assessments. ^{Reference Jeunet, Jahanpour and Lotte19} For example, in a recent study, participants played a gamified “stunt plane” task with adaptive difficulty, and optimal engagement (measured via EEG indices like β/(θ + α)) correlated with high classification accuracy (within-subject F1 ≈ 0.89, cross-subject F1 ≈ 0.81), indicating that sustained engagement yields cleaner, more classifiable EEG patterns. ^{Reference Ahmed, Ferguson-Pell, Adams and Ríos Rincón84} These findings point to adaptive difficulty as a particularly valuable design feature, one that simultaneously sustains user motivation and stabilizes EEG signal quality, making it a practical target for future BCI game development protocols.

BCI technologies used to assess cognition, the electrode placement used and the brain waves explored

The review revealed considerable variability in the BCI technologies and setups used to assess cognition. While the 10–20 system was the standard for electrode placement, the number of channels varied widely, ranging from single-channel consumer-grade devices to high-density 64-channel systems, reflecting a tension between accessibility and signal quality. Lower-cost, widely used devices like the MindWave and Emotiv EPOC+, the most frequently reported technologies in this review, offer ease of use and portability but are limited by factors such as fewer channels, fixed electrode positions and lower signal fidelity. ^{Reference Rieiro, Diaz-Piedra and Morales85,Reference Williams, McArthur, de Wit, Ibrahim and Badcock86} Reduced electrode counts and fixed placements can affect both sensitivity and specificity. With fewer electrodes, the system has reduced spatial coverage, which limits its ability to detect subtle or localized neural signals associated with specific cognitive processes, thereby reducing sensitivity. Cognitive assessment research confirms that higher channel counts yield better sensitivity. ^{Reference Shah, Golmohammadi, Ziyabari, Von Weltin, Obeid and Picone87} Specificity is also affected, as narrow spatial coverage and the inability to place electrodes optimally make it difficult to distinguish between true cognitive events and non-target or artifact activity. Previous research has demonstrated that reducing channel count is associated with increased false positive rates, suggesting that consumer-grade devices, ranging from a single channel (MindWave) to 14 channels (Emotiv EPOC+), may compromise the accuracy and reliability of cognitive signal detection despite their accessibility and ease of use. ^{Reference Shah, Golmohammadi, Ziyabari, Von Weltin, Obeid and Picone87,Reference Song, Jeong, Kim, Jang and Kim88} Research-grade systems like BrainProducts ActiChamp, with 32–64 or more channels, provide higher resolution and more precise data and yet remain less feasible for scalable or real-world applications due to cost and complexity. Future studies should directly compare consumer-grade and research-grade systems within the same paradigm to quantify the trade-off between accessibility and signal fidelity and to establish the minimum channel count required for reliable cognitive assessment.

The choice of technology influences both the cognitive domains assessed and the brain signals captured. For instance, many studies using MindWave and Emotiv relied on predefined metrics like “attention“ and “meditation” scores based on proprietary algorithms, which can obscure the interpretability and reproducibility of results. These devices tend to focus on frontal or central sites (e.g., Fz, Cz, Pz), aligning with their emphasis on attention and executive functions. Beta (β) waves were the most frequently analyzed, followed by alpha (α) and theta (θ), reflecting their established relevance in cognitive tasks such as attention modulation, working memory and task engagement. ^{Reference Spitzer and Haegens56–Reference Klimesch58} In contrast, delta (δ) and gamma (γ) activity were reported in only one-third of the studies, suggesting that cognitive processes linked to deep learning states or higher-order integrative functions may be underrepresented in current BCI game research. ^{Reference Harmony61,Reference Fries89}

Beyond device selection, the field would benefit considerably from standardized reporting practices across all stages of EEG data collection and analysis. The studies included in this review varied widely in how they reported electrode placement, sampling rates, preprocessing steps and analytic pipelines, making cross-study comparison difficult and limiting reproducibility. Inconsistent reporting of preprocessing decisions, such as filtering parameters, artifact rejection methods and reference schemes, has been shown to systematically alter signal quality and lead to notably different conclusions even when the same paradigm is used. ^{Reference Vos, Ebrahimpour, van Eijk, Sarnyai and Rahimi Azghadi90} Adopting transparent and consistent reporting standards, including full documentation of electrode configurations, sampling rates, artifact-handling procedures and feature-extraction methods, is essential for building a cumulative and reproducible evidence base in BCI game research. Frameworks and open-source pipelines, such as Pipeline for EEG Processing and EEGLAB-based toolchains, have been proposed to support this standardization effort and should be considered reference points for future BCI game studies. ^{Reference Bigdely-Shamlo, Mullen, Kothe, Su and Robbins91,Reference Delorme and Makeig92}

As BCIs are increasingly considered for use with vulnerable populations, ethical and accessibility considerations must be addressed. Neural data collected through EEG can reveal sensitive information about cognitive and emotional states, raising concerns around data privacy and informed consent, particularly when participants are children or individuals with cognitive, speech and/or motor impairments. ^{Reference Burwell, Sample and Racine93} Physical comfort and wearability of EEG devices also present practical challenges for these groups, as device fit and setup time can affect both participant tolerance and data quality. ^{Reference Huggins, Karlsson and Warschausky17} Future studies should adopt user-centered design principles and clear data governance protocols to ensure BCI games are both ethically sound and genuinely accessible.

Limitations

This scoping review has limitations that should be acknowledged. First, the review included only English-language publications, which may have introduced language bias and excluded potentially relevant studies published in other languages. ^{Reference Dickersin, Min and Meinert94} Furthermore, the exclusion of documents, such as theses and book chapters, may have led to an underrepresentation of emerging research in this field. Second, a device-related bias should be noted: several studies relied on proprietary metrics such as the “attention” and “meditation” scores generated by consumer-grade devices like MindWave and Emotiv EPOC+. Because these metrics are derived from undisclosed algorithms, they limit cross-study comparability and make it difficult to draw consistent conclusions about the cognitive constructs being assessed. Third, as is consistent with scoping review methodology, no formal quality appraisal of included studies was conducted; while this is appropriate given the review’s aim to map the breadth of available evidence, it means that study quality was not systematically weighted, and publication bias, whereby studies with positive or significant findings are more likely to be published, cannot be ruled out.

Future direction

Future research should address several critical priorities. First, BCI game-based cognitive assessments must be expanded to include diverse and underrepresented populations, particularly children with DD beyond ADHD, older adults with conditions like dementia and individuals with motor or communication impairments who could benefit most from non-verbal assessment tools. Second, the range of cognitive domains assessed should extend beyond attention to include memory, problem-solving, executive function and cognitive flexibility, through the development and validation of task-specific games. Third, future studies should rigorously validate BCI-based assessments against gold-standard neuropsychological measures using robust psychometric frameworks to establish reliability, validity and sensitivity to cognitive change. Fourth, transparent and consistent reporting of EEG configurations, preprocessing steps and analytic pipelines is essential to improve reproducibility and cross-study comparability. Finally, larger, longitudinal and ecologically valid studies are needed to assess the feasibility of integrating BCI games into clinical and educational workflows, examining outcomes such as user engagement, cognitive development trajectories and real-world diagnostic utility.