Go and the Brain: Cognitive and Neural Impacts of Training
Article information
Abstract
Objective
This review synthesizes evidence on the cognitive and neural impacts of the strategic board game Go, emphasizing its role in enhancing cognitive functions and inducing neuroplastic changes.
Methods
We analyzed studies investigating the effects of Go on diverse populations, including novices, professional players, and older adults. Research included neuropsychological assessments, neuroimaging findings (fMRI, PET), and behavioral outcomes from randomized controlled trials (RCTs).
Results
Go training enhances executive functions, memory, and visuospatial reasoning, as demonstrated by studies on novices and clinical populations. Professional Go players show domain-specific neural adaptations, including increased precuneus and cerebellar activation during decision-making tasks. In clinical settings, Go interventions improve mood, reduce anxiety, and increase neurotrophic factors like brain-derived neurotrophic factor.
Conclusion
Go training demonstrates significant potential as a cognitive intervention to promote brain health, emotional well-being, and resilience against cognitive decline. Further longitudinal studies are required to validate its long-term effects.
INTRODUCTION
Cognitive decline represents a significant challenge for aging populations, underscoring the necessity for efficacious strategies to preserve brain health and cognitive function. One such potential intervention is the board game Go, also known as Baduk in Korea. The game has a long history and is of significant cultural importance. Go, originating in China over four thousand years ago, has spread across East Asia and the globe, becoming one of the longest-standing strategy games in existence. In the 20th century, Go experienced a surge in popularity in Western countries, particularly following the Second World War. Currently, Go is played by millions of people around the globe, with South Korea alone having over ACCESS10 million enthusiasts. The advent of online platforms has further facilitated access to Go, enabling players to engage in competition and collaboration with one another irrespective of geographical boundaries [1-5].
The advent of artificial intelligence (AI) has had a profound impact on the game of Go. In 2016, the AI program AlphaGo, developed by Google DeepMind, achieved a decisive victory against several of the world’s top human players, including South Korean champion Lee Sedol. This seminal accomplishment exemplified the capacity of AI to excel in intricate strategic pursuits, thereby propelling Go to global prominence. Furthermore, AlphaGo’s innovative strategies have influenced human players, prompting a re-evaluation of Go theory and inspiring a new era of creativity in the game. AI tools are now commonly employed for training and analysis, assisting players in enhancing their abilities and grasping the nuances of the game to an unparalleled degree [6,7]. A recent study demonstrated that superhuman AI can encourage novel decision-making by humans in certain domains, which suggest that innovative thinking can spread from machines to humans and among humans themselves, possibly improving human decision-making in those domains [8].
The game’s straightforward rules belie its profound intricacies, necessitating players to engage in sophisticated decisionmaking, strategic planning, and pattern recognition, thereby exemplifying an epitome of intellectual challenge and strategic depth. Go is distinctive among other cognitive training tools due to its distinctive combination of visuospatial demands, adaptive strategies, and emotional engagement. A growing body of research indicates that Go engages a multitude of brain regions, facilitates social interaction, and cultivates competencies that are essential for daily life, including focus, resilience, and problem-solving abilities [9-13].
This review examines the cognitive and neural impacts of Go, emphasizing its potential as an intervention for cognitive enhancement, emotional well-being, and resilience against aging-related cognitive decline. This review synthesizes findings from neuropsychological assessments, neuroimaging studies, and clinical trials with the aim of providing a comprehensive understanding of how Go influences brain function and why it is particularly suited for addressing contemporary challenges in cognitive health. If Go is effective in enhancing cognition and brain health in late life, it will be an effective activity for promoting cognitive health in the elderly. This is because it is already a popular pursuit, relatively simple to learn, and can be played online with anyone in the world or superhuman AI without any space or time constraints, which makes it an easily implementable intervention.
METHODS
This review is a selective narrative review aimed at synthesizing and critically analyzing the cognitive and neural impacts of Go. The narrative approach was chosen to integrate findings from diverse methodologies, including cross-sectional, observational, and randomized controlled trials (RCTs), rather than restricting the scope to longitudinal studies alone. Unlike systematic reviews, which focus exclusively on specific research questions, the narrative format accommodates a broader perspective, synthesizing evidence from multiple methodologies to understand both long-term and shortterm effects of Go training. This decision allows for a broader understanding of the current evidence while identifying gaps that future research can address.
A systematic approach was employed to gather and analyze existing studies on the cognitive and neural impacts of playing Go. To ensure comprehensive and unbiased coverage, a multitude of electronic databases were utilized, including PubMed, Scopus, and Web of Science. The search strategy was specifically designed to identify studies related to the cognitive and neural outcomes of playing Go, excluding irrelevant or tangential topics. The search terms and filters applied included “Go game” or “baduk” in titles or abstracts, while excluding references to “Pokemon,” “No Go,” and “AI.” The searches were conducted in English, without restrictions on publication dates, to encompass both historical and contemporary research contributions. The search terms used were: (“Go game” [Title/Abstract] OR baduk [Title/Abstract] NOT Pokemon [Title/Abstract] NOT “No Go” [Title/Abstract] NOT AI [Title/Abstract] AND English [lang] for PubMed; TITLE-ABS-KEY [“Go game” OR baduk] AND NOT TITLE-ABS-KEY [Pokemon]) AND NOT TITLEABS-KEY (“No Go”) AND NOT TITLE-ABS-KEY(AI) AND LANGUAGE (English) for Scopus; TS=(“Go game” OR baduk) NOT TS=(“Pokemon”) NOT TS=(“No Go”) NOT TS=(“AI”) AND LA=(“English”) for Web of Science.
To ensure rigor in the selection process, studies were included based on the following criteria, emphasizing the inclusion of studies focusing solely on Go and the exclusion of those related to other cognitive training methods: 1) the study must focus on cognitive or neural outcomes related to playing Go; 2) the study must involve human participants; 3) the study must be published in peer-reviewed journals; 4) the study must utilize cross-sectional, observational, RCT, or longitudinal designs relevant to the research objectives; 5) studies with non-peer-reviewed or speculative content, such as conference abstracts, were excluded. While longitudinal studies provide the strongest evidence for training effects, this review also includes cross-sectional, observational, and RCTs to provide a broader understanding of the cognitive and neural implications of Go. This approach allows for an examination of both long-term adaptations in experienced players and the immediate effects of Go training interventions. Additionally, non-English studies were included only if they had translated abstracts that allowed for full-text review. This approach was adopted to facilitate a more inclusive review while ensuring the validity and relevance of the selected studies. Studies were excluded if they primarily focused on AI applications of Go, discussed unrelated topics such as gaming culture or social aspects of Go, or were non-peer-reviewed articles or conference abstracts. For each study deemed eligible, key information was extracted, including sample size, participant demographics, study design, cognitive or neural outcomes, and statistical findings, such as effect sizes and significance levels. Due to the heterogeneity in study designs, outcome measures, and participant characteristics across the included studies, meta-analytical techniques were not applied.
The database searches yielded 15 studies from PubMed, of which 11 met the inclusion criteria following abstract and full-text review. Scopus yielded 160 results, of which 11 studies satisfied the inclusion criteria. Web of Science identified 93 studies, of which 12 were deemed relevant for the review. After removing duplicates across the three databases, a total of 12 unique studies were identified (Table 1). These 12 studies were utilized for the review, representing a diverse array of methodologies and outcomes, and providing a comprehensive overview of the current understanding of the cognitive and neural impacts of playing Go.
A summary of the key studies
In this review, we categorize the effects of Go training into short-term and long-term based on the duration of engagement: 1) long-term effects: these pertain to cognitive and neural adaptations identified in individuals with extensive Go experience spanning several decades, such as professional Go players who have dedicated a significant portion of their lives to the game; 2) short-term effects: these refer to cognitive and neural changes observed in individuals who have participated in Go training programs lasting from several weeks up to several months.
Short-term effects: these refer to cognitive and neural changes observed in individuals who have participated in Go training programs lasting less than one year.
1. Long-term effects: effects associated with Go activities extending decades, including data from professional Go players who have engaged in the game throughout their lives.
2. Short-term effects: effects observed from Go training programs lasting less than one year.
Notwithstanding the systematic approach that was employed, it is imperative to acknowledge the limitations that are inherent to the study. The focus on English-language publications may have resulted in the exclusion of studies published exclusively in other languages without translated abstracts. Additionally, the heterogeneity of study designs limited the feasibility of meta-analytical synthesis, necessitating a narrative approach to data integration and interpretation. Future reviews could benefit from broader linguistic inclusion and the development of standardized methodologies to facilitate quantitative comparisons.
RESULTS
Long-term effects of playing Go
A total of seven studies have previously investigated this topic, and the results of these studies consistently demonstrate significant changes in cognitive abilities, functional connectivity, and brain structure among Go experts compared to novices.
Research on the cognitive performance and neural activity of Go experts compared to novices has demonstrated marked differences in visuospatial abilities, pattern recognition, and decision-making processes. In a study involving 60 male participants, Go experts demonstrated significant superiority in visuospatial tasks and pattern recognition tests, achieving an accuracy of 76.97% compared to 65.08% for novices [14]. These findings suggest that long-term Go training enhances domain-specific and general cognitive skills. Neurofunctional studies reveal distinct activity patterns between Go experts and amateurs during strategic decision-making tasks. For instance, during life-or-death judgment tasks, Go experts exhibited significant activation in the precuneus and cerebellum, while amateurs demonstrated broader activation in the premotor and parieto-occipital cortices [15]. Notably, the right cerebellum activity in experts exhibited a positive correlation with task performance (r=0.61), underscoring its involvement in strategic thinking. These regions are critical for visuospatial processing and strategic thinking, suggesting that long-term engagement with Go enhances these cognitive domains. The heightened activity in these areas implies a greater reliance on specialized neural pathways, leading to improved efficiency in processing complex board positions and anticipating opponent moves.
Resting-state functional MRI studies have elucidated the enhanced efficiency of neural networks in Go experts. Increased global efficiency and decreased characteristic path lengths have been reported in functional brain networks, suggesting optimized information processing [16]. These alterations in functional connectivity may contribute to enhanced cognitive efficiency, allowing players to integrate visuospatial and strategic information more effectively. Furthermore, experts demonstrated heightened extrinsic connectivity between regions that extend beyond conventional networks, such as the default mode network and the salience network, in comparison to novices. These findings indicate that Go training may facilitate complex cognitive processes by improving attentional control and decision-making flexibility during strategic tasks [17].
Structural neuroimaging studies employing voxel-based morphometry and diffusion tensor imaging further reveal significant differences in brain morphology between Go experts and novices. Experts demonstrated heightened gray matter volume (GMV) in regions implicated in reward and decision-making, such as the bilateral nucleus accumbens, while exhibiting diminished GMV in the right amygdala [16]. Furthermore, decreased GMV in the medial orbitofrontal cortex exhibited a negative correlation with years of Go training (r=-0.802), suggesting a neural adaptation to the demands of long-term strategy-based training. White matter integrity studies revealed increased fractional anisotropy (FA) in the frontal lobe, cingulum, and striato-thalamic regions, indicative of enhanced connectivity that supports executive functions, attention, and problem-solving [18]. Conversely, decreased FA in the premotor and parietal regions suggests reduced reliance on working memory, suggesting a shift toward memory chunking and efficient neural processing strategies. These structural adaptations reinforce the notion that Go training induces long-term neural plasticity, leading to cognitive improvements that extend beyond the game itself.
The integration of findings from functional and structural neuroimaging studies offers a comprehensive understanding of the neural mechanisms underlying the cognitive benefits of long-term Go training. The observed increases in precuneus and cerebellar activity highlight their roles in visuospatial reasoning and motor planning, while connectivity changes point to enhanced cognitive efficiency. Increased precuneus and cerebellar activity, indicative of reliance on visual imagery and motor planning, is associated with functional differences. In contrast, structural differences, including increased GMV in the nucleus accumbens and FA in strategic networks, support these cognitive functions at a morphological level. This alignment suggests that long-term Go training induces plastic changes in both gray and white matter, facilitating the development of efficient neural pathways for visuospatial processing, strategic planning, and decision-making. The observed reductions in GMV and FA in certain regions, such as the medial orbitofrontal cortex and premotor areas, reflect neural optimization, as players transition from reliance on generalized cognitive functions to specialized and efficient strategies.
This comprehensive analysis underscores the substantial impact of Go on the brain’s functional and structural organization, reinforcing the pivotal role of long-term training in enhancing neuroplasticity and cognitive performance. Future research endeavors should aim to elucidate how these observed adaptations compare to those induced by other cognitively demanding activities. This comparative analysis will provide a more nuanced understanding of Go’s unique contributions to brain and cognitive health.
Short-term effects of playing Go
A limited number of studies have investigated the shortterm effects of Go interventions, particularly in clinical populations such as children with attention deficit hyperactivity disorder (ADHD), older adults with mild cognitive impairment (MCI), and patients with Alzheimer’s disease (AD). These studies provide evidence for cognitive and emotional benefits, supported by neuroimaging findings that reveal neural changes associated with Go training.
Clinical trials evaluating short-term Go interventions consistently report significant improvements in cognitive functions and emotional well-being across various populations. For instance, a 16-week Go training program administered to 17 children diagnosed with ADHD led to a reduction in inattention scores (z=2.94) and an increase in forward digit span scores (z=2.21). These findings suggest enhanced working memory and executive function as a result of Go-based interventions [11]. In older adults at risk of cognitive decline, a 12-week Go intervention improved visual memory span test scores. The face-to-face sessions demonstrated a substantial effect size (Cohen’s d=0.89), while the tablet-based sessions exhibited a medium effect size (Cohen’s d=0.67) [19]. Furthermore, the efficacy of Go training in preserving cognitive functions, such as attention and working memory, was substantiated by consistent or enhanced performance on backward digit span tasks in experimental subjects when compared to a decline observed in control groups. The benefits of short-term Go interventions extend to emotional well-being. AD patients who engaged in daily Go sessions for a period of 6 months exhibited substantial reductions in depression (Montgomery-Åsberg Depression Rating Scale: -4.72, 95% confidence interval [CI], 0.69–9.12) and anxiety (Hospital Anxiety and Depression Scale: -1.75, 95% CI, 0.17–3.68), alongside improvements in quality of life scores (RAND 36-Item Health Survey: +4.61, 95% CI, -2.75–11.32) [12].
Go interventions demonstrated remarkable adaptability across varying levels of cognitive impairment. In a study involving older adults, all participants understood the basic rules of Go, and most individuals with MCI or mild dementia (CDR 1) successfully played the game. Additionally, some participants with moderate dementia (CDR 2) were able to learn and execute advanced gameplay techniques, indicating that Go retains potential as a cognitive training tool even in populations with significant cognitive challenges [19].
Functional neuroimaging studies using 18F-FDG PET provide insights into the neural changes associated with shortterm Go interventions. In a study of older adults, participants in the intervention group showed increased glucose metabolism in the left middle temporal gyrus and bilateral putamen (p<0.01), brain regions associated with semantic memory and motor planning. These metabolic changes were found to correlate positively with Go skill improvement scores (r=0.512, p<0.05) [20]. In contrast, the control group, which participated in health education lectures, exhibited increased activity only in the left superior frontal gyrus, reflecting less targeted cognitive engagement. These findings suggest that Go training specifically enhances neural activity in regions crucial for learning and executing strategic tasks. Furthermore, the correlation between metabolic changes and Go skill acquisition highlights the intervention’s potential to foster rapid neural adaptation in cognitive and motor domains.
The benefits of short-term Go interventions vary across populations. For children with ADHD, the focus on executive function and attention aligns well with their cognitive needs, making Go a suitable therapeutic activity. In contrast, the improvements in memory and emotional stability observed in older adults suggest a broader applicability for delaying or mitigating cognitive decline. Moreover, AD patients not only demonstrated cognitive gains but also increased serum brainderived neurotrophic factor levels (p<0.001), which are linked to neuroprotective effect [12]. However, short-term interventions have limitations, including variability in effect sizes depending on intervention mode (face-to-face vs. tablet-based) and individual differences in engagement levels. Moreover, the lack of long-term follow-up in these studies makes it challenging to assess the durability of observed benefits.
The findings from these studies illustrate the potential of Go as an effective short-term cognitive intervention. The neuroimaging evidence of increased metabolic activity in key brain regions supports the functional improvements in executive function, memory, and emotional stability observed in the studies. These results suggest that Go can foster rapid cognitive and neural adaptations in a short time frame, making it a versatile tool for targeted interventions. Further studies are needed to determine the optimal duration, frequency, and mode of Go interventions to maximize their effectiveness across diverse populations. Furthermore, comparative analyses with other cognitive activities could elucidate the distinctive contributions of Go as a short-term therapeutic strategy. While the majority of studies have emphasized the cognitive and neural benefits of Go, one study reported the occurrence of reflex epilepsy in a small subset of patients who experienced seizures triggered by playing Go or related games [21]. This finding underscores the necessity for meticulous consideration of individual neurological vulnerabilities when selecting participants for Go-based interventions. These cases suggest that pre-screening for susceptibility to reflex epilepsy may be imperative to ensure the safety of participants, particularly in populations with known neurological risks.
DISCUSSION
Long-term and short-term Go interventions have been shown to have significant impacts on brain function and cognitive abilities, albeit with distinct characteristics. Shared benefits include enhancements in working memory, attention, and strategic decision-making. Neuroimaging findings from both approaches highlight increased activity in key regions such as the left middle temporal gyrus and bilateral putamen, which are integral to semantic memory, motor planning, and strategic learning. These shared effects suggest that Go engages fundamental neural mechanisms involved in complex cognitive tasks. The adaptability of Go as a cognitive intervention is particularly notable in populations with cognitive impairments, such as MCI and dementia. The ability of participants across cognitive stages to engage with Go highlights the game’s potential as a versatile and accessible tool for cognitive stimulation. For individuals with dementia, Go can serve as an engaging activity that not only maintains cognitive functions and emotional stability but also offers opportunities for cognitive growth (Table 2).
A summary on the effects of Go game
Long-term Go training, however, uniquely induces structural changes in the brain, including increased GMV in the bilateral nucleus accumbens and enhanced white matter integrity in the frontal and striato-thalamic regions. These adaptations are associated with improved global network efficiency and optimized neural communication pathways, as evidenced by resting-state functional MRI studies. These structural changes support higher-order cognitive functions, such as intuitive decision-making, visuospatial processing, and the development of task-specific neural templates. By contrast, short-term interventions primarily yield rapid, domain-specific improvements, such as enhanced attention and emotional regulation. These effects are particularly pronounced in clinical populations, including children with ADHD and older adults with MCI, where interventions have led to reductions in inattention scores, improved working memory, and alleviation of depression and anxiety symptoms. The emotional benefits, such as reductions in anxiety and improvements in quality of life, appear more robust in short-term interventions, likely due to immediate effects on brain metabolism and neurotransmitter systems. A comparison of the findings reveals that long-term training offers more enduring and structural brain changes, while short-term interventions deliver immediate and context-specific benefits. Together, these results suggest that Go could serve as both a preventive measure for cognitive decline and a therapeutic tool for addressing specific cognitive deficits. The integration of long-term training with short-term interventions may offer a comprehensive approach to cognitive and neural enhancement, leveraging the strengths of each method to achieve both immediate and sustained benefits.
The current body of research on Go’s cognitive and neural effects spans a wide age range, from adolescents to older adults. However, studies encompassing both young and elderly participants are scarce, and none have directly examined how age influences the benefits derived from Go practice. Given the well-established relationship between age and cognitive function, it is plausible that the impact of Go varies across different age groups. For instance, younger individuals may experience enhancements in executive functions and attention, while older adults might benefit from maintenance or improvement in cognitive flexibility and memory. Consequently, tailoring Go training programs to specific age groups could optimize their effectiveness. Future research should focus on systematically investigating the interaction between age and the cognitive benefits of Go, potentially leading to agespecific training protocols that maximize the game’s positive effects on the brain.
Notwithstanding the encouraging initial findings, the extant body of research is encumbered by salient limitations that demand consideration. First, the predominance of crosssectional study designs, which hinder the ability to establish causal relationships between Go training and observed cognitive and neural changes. Longitudinal studies are necessary to track changes over time and confirm whether the observed effects persist beyond the training period. Second, the lack of gender diversity in participant samples limits the generalizability of the findings. Many studies on Go training have predominantly focused on male participants. Future research should aim to include more diverse participant pools to assess potential gender differences in cognitive and neural adaptations resulting from Go training. Third, cultural factors also play a crucial role in the impact of Go on cognitive function. As Go is more widely practiced in East Asian countries, many studies have been conducted within these populations, potentially limiting the generalizability of the findings to individuals from different cultural backgrounds. Investigating the effects of Go training in populations with varied cultural and educational experiences could provide a more comprehensive understanding of its cognitive benefits. Fourth, there is a pressing need for more RCTs with larger and more diverse cohorts. While existing RCTs provide initial evidence for the cognitive benefits of Go, their sample sizes are often small, and intervention durations vary significantly. Conducting well-powered RCTs with standardized training protocols and longer follow-up periods will help establish a stronger empirical foundation for Go as a cognitive training tool. Fifth, the specificity of the tasks employed in these studies gives rise to questions regarding the uniqueness of the observed effects to Go and their potential generalization to other cognitively demanding activities. Sixth, variability in delivery methods, such as face-to-face versus tablet-based training, further complicates standardization and comparison across studies. Seventh, there is a paucity of studies that have thoroughly examined the role of individual differences, such as baseline cognitive abilities or neurological vulnerabilities, in determining intervention outcomes. Finally, a particularly salient limitation in dementia-related studies is the lack of exploration into the factors that facilitate engagement and success at different stages of cognitive impairment. While participants across varying levels of dementia have been shown to engage with and benefit from Go, future research should investigate how elements like the duration and intensity of training, instructor support, and tailored game modifications influence outcomes. By elucidating these dynamics, Go interventions can be optimized to maximize their impact for diverse cognitive profiles, ensuring both efficacy and inclusivity.
Although this study a selective narrative review, the inclusion of key study selection details directly in the Methods section strengthens the transparency of this review. By clearly documenting the number of studies identified, screened, excluded, and included, this approach ensures methodological clarity. Additionally, discussing reasons for study exclusion and outlining inclusion criteria provides readers with a comprehensive understanding of the review process. These steps highlight the rigor of the selective narrative approach while maintaining accessibility and flexibility in synthesizing diverse evidence on Go training.
Future research should solidify the role of Go in neuropsychiatric and cognitive interventions by addressing current research limitations. Longitudinal studies should be conducted to establish causal links between Go training and observed changes in brain and cognitive functions in the study samples with enhanced diversity. There is a pressing need for more RCTs with larger and more diverse cohorts. While existing RCTs provide initial evidence for the cognitive benefits of Go, their sample sizes are often small, and intervention durations vary significantly. Conducting well-powered RCTs with standardized training protocols and longer follow-up periods will help establish a stronger empirical foundation for Go as a cognitive training tool. In addition, comparative studies that explore the effects of Go training in relation to other cognitive activities, such as chess or language learning, could provide insights into Go’s unique contributions to brain and cognitive health. The development of integrated interventions that combine the immediate benefits of short-term training with the sustained advantages of long-term practice holds promise in optimizing outcomes for diverse populations. Mechanistic studies employing advanced neuroimaging and biomarker analysis hold promise in further elucidating the neural and molecular underpinnings of Go’s effects.
In conclusion, Go training demonstrates substantial potential as both a long-term tool for enhancing neuroplasticity and higher-order cognitive functions and a short-term intervention for improving attention, memory, and emotional well-being. Addressing the current research limitations and exploring its long-term impacts will solidify its role in neuropsychiatric and cognitive interventions.
Notes
Availability of Data and Material
Data sharing not applicable to this article as no datasets were generated or analyzed during the study.
Conflicts of Interest
The author has no potential conflicts of interest to disclose.
Funding Statement
This study was supported by a grant from the Korea Baduk Association (grant no. 06-2024-0330).
Acknowledgments
None