Abstract
Background:
Traditional board games can entail significant skills encompassing several cognitive functions across different domains. Therefore, they may potentially represent effective cognitive interventions in the aging population with or without Alzheimer’s disease or other types of dementia.
Objective:
We aimed at verifying the hypothesis that traditional board games can prevent or slow down cognitive decline, through a systematic review on traditional board games and dementia.
Methods:
We searched five databases with tailored search strings. We included studies assessing the impact of board games on elderly subjects at risk of or suffering from cognitive impairment, or subjects with cognitive impairment irrespective of age. Studies where the effect of board games was not separated by cards or other games were excluded. A meta-analysis was performed for specific cognitive and non-cognitive outcomes.
Results:
Board games improved mental function, as measured by Montreal Cognitive Assessment (p = 0.003) and Mini-Mental State Examination (p = 0.02). Ska and Go improved Trail Making Test –A, while Mahjong improved executive functions. There was no consistent effect across different games on Digit Span or Categorical Fluency. Chess improved quality of life measured with the WHO-QoL-OLD scale (p < 0.00001). Mahjong temporarily improved depressive symptoms. Go increased BDNF levels and left middle temporal gyrus and bilateral putamen metabolism.
Conclusions:
Traditional board games may slow global cognitive decline and improve the quality of life in elderly subjects. Different games have varying impacts on specific cognitive domains, possibly mediated by functional and biological factors.
Keywords
“I’m good enough,” the King said, ‘only I’m not strong enough. You see, a minute goes by so fearfully quick. You might as well try to stop a Bandersnatch!”
Lewis Carroll, Through the Looking-Glass –Chapter 7
INTRODUCTION
Board games have been played by humans for millennia. Historically, a significant proportion of them evolved to allow a player to use different pieces or “men” to recreate a relevant figurative scenario, such as a battle, a hunt, or a race. The way these pieces are used, the objectives, and the rules allow board games to be divided into games of chance and games of skill, with some games showing aspects of both categories [1]. Games of skill may be extremely complicated, either because of their rules, their strategies, and tactics, or all. A typical well-known example in the Western world is represented by chess, which is notoriously highly cognitively demanding [2–4]. A number of cognitive domains influence performance in board games, most notably attention, perception, executive functions, visuospatial abilities, and even memory, each in various degrees according to the player’s level [5]. Another typical element of board games is their inherent social component, since they are typically played against other humans. Moreover, board games are an integral part of cultural identity. The neural basis underneath board game mastering has received significant attention as well, and research often points at brain structures that represent disease epicenters (e.g., precuneus), thus potentially implying interesting cognitive reserve mechanisms [6, 7].
For all these reasons, board games can be regarded as a potential multidomain cognitive intervention in people at risk of or already suffering from cognitive decline. A few prospective studies have shown that playing games seem to lower the risk of incident dementia and even depression, while cross-sectional studies have shown an association between playing games and lower prevalent dementia. In many of these studies, the definition of the specific games was quite broad, including also activities such as bingo or card games, which generally rely more on chance than traditional board games of skill [8–13]. In other cases, the type of games was not even specified [14, 15]. In these cases, it is possible that what is measured was not the effect of cognitive intervention, but rather a product of social activities, or even a result of reverse causality, with people at risk of cognitive impairment being less willing and/or able to play cognitively demanding games.
In other cases, specifically designed or modern commercial board games were used as a cognitive intervention in elderly subjects [16, 17]. While the reported effect is generally positive, it is difficult to imagine that such games may be implemented as a large-scale cognitive intervention, because of both cultural and economic factors. On the contrary, traditional board games such as chess or Go are easily available, almost inexpensive (especially today, with the availability of many free apps), culturally acceptable, and shared by large communities. There are gross regional differences in the predominant game in specific cultures, with chess being more popular in the Western world, Go being more diffused in China, Korea, and Japan, and mancala-type games being prevalent in Africa and South-Eastern Asia. Other games such as checkers or backgammon are less played, and mostly in Europe [1].
Our aim was to evaluate the impact of such traditional board games on cognitive decline in the elderly, and their potential to prevent or slow down dementia. Therefore, we conducted a systematic review of studies investigating traditional board games in the elderly and provided a quantitative summary of the available evidence.
METHODS
This systematic review has been conducted in accordance with the principles of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement.
We conducted electronic searches for eligible studies within each of the following databases, up to March 4, 2023: Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE (accessed via PubMed), Embase, Scopus, and Web of Science. The reference list of all screened studies was checked to identify further eligible studies. Records were identified with tailored search strings for each database. These can be found in the Supplementary Material. Given the questionable nature of mahjong, which could be interpreted both as a board game and as a card game, it was not included in the first search strings. A sensitivity search on PubMed with the string Mahjong AND Dementia yielded only 10 results; of these, three were already present in our search, two were screened, and the remaining five were excluded. Of note, in two of these studies, the authors classified mahjong as card games. However, we applied the same inclusion and exclusion criteria reported below for mahjong studies, therefore considering it as a board game for the purposes of this review.
Citations identified from the literature searches were imported to Mendeley and duplicates were automatically removed by the software. Then, two reviewers (FEP and LT) independently screened the titles and abstracts of all the records to match the inclusion criteria. A consensus was reached through discussion in case of disagreements. The full texts of all eligible studies, reviews, and meta-analyses were retrieved, and their reference lists were checked.
The inclusion criteria were as follows: Studies on traditional abstract board games (i.e., played since at least the XIX century) and cognitive impairment; Studies on human subjects, including cohort studies (prospective or retrospective), cross-sectional studies, and randomized and non-randomized controlled trials; Studies on elderly patients (age > 60 as per WHO standards), with or without cognitive impairment, or on patients with cognitive impairment irrespective of age Outcomes: at least one cognitive outcome (neuropsychological tests, prevalent or incident dementia) Studies on at least 10 subjects; Full text in English available
The exclusion criteria were as follows: Studies on newly developed ad hoc board games, cards, or other kinds of games; Studies in which the effect of board games was not analyzed separately from other kinds of games; Studies evaluating cognitive effect of board games in cognitively normal middle-aged subjects (mean age < 60); Review articles or meta-analyses, case reports, case series, cross-sectional studies, conference abstracts, letters, and editorials.
We chose to limit our analysis to broadly available and diffused board games, in order to maximize the potential implementation of a cognitive intervention with them. We decided to include articles in which the specific type of board game was not reported, but they were nevertheless clearly separated from card games or other types of games. Review and meta-analysis reference lists were checked to identify additional studies and to elucidate theoretical aspects of the discussion.
Data extraction was performed by one reviewer (FEP), and all proceedings were checked by a second reviewer (LT). We extracted data about authors, year, country, number of participants, diagnosis, setting, type of game, type of study, duration, follow-up, dropouts, and results.
Risk of bias assessment
The risk of bias was assessed only for randomized and non-randomized controlled studies, with the RoB 2 and the ROBINS-I Cochrane tools respectively. Cluster-randomized trials were appraised with the RoB2 CRT tool. Bias assessment was performed independently by two reviewers (FEP and LT), and disagreements were resolved by consensus.
Data analysis
We could only perform a meta-analysis on nine outcomes, by pooling data of change from baseline for different games together. For each outcome, we calculated the distributions of changes for intervention and controls with the following formulas:
To account for the heterogeneity of the studies, we used a random effect model and mean difference or standardized mean difference depending on the fact that the outcome was assessed with the same units or with different units. A sensitivity analysis with a fixed effect model did not change the results (data not shown). The meta-analyses were performed using the RevMan 5.4.1 software, available at https://training.cochrane.org/online-learning/core-software/revman/revman-5-download. Whenever possible, we also reported effect sizes as Cohen’s d for individual studies, following previously published methods [18]. The other results of the current review are presented narratively.
RESULTS
We identified 2,902 records through the initial search. After duplicate removal, 2,772 titles were screened according to title and abstract, of which 2,723 were excluded. For three studies, the full text was not available. Of the 46 studies assessed for eligibility, nine were included in this review. We added other six reports after checking the reference lists of the retrieved studies. Finally, 15 studies were included in this review (the flow diagram is shown in Fig. 1).
PRISMA flow-chart.
A list of the relevant data from each study is presented in Tables 1 2. Only four specific board games have been assessed as potential cognitive interventions: chess (three articles), Go (four articles), mahjong (seven articles), and ska (one article). In one article, the effect of chess was not separated from the effect of mahjong. In another further article, the specific type of board game was not specified. In the following sections, the results will be divided by game. Each section will begin with a summary of the game rules, assuming that the reader may not be familiar with them.
Characteristics of RCT and nRCT
Final participants per groups are listed in columns labeled with n (drop-outs are in brackets). Duration of the intervention and total follow-up are in months. AD, Alzheimer’s disease; C, community; CDR-SB, Clinical Dementia Rating-sum of boxes; DC, day care; DS, Digit Span; DSB, Digit Span Backward; DSF, Digit Span Forward; DST, Digit Span Test; FAQ, functional activities of daily living; GAF, Global Assessment of Functioning; GDS, Geriatric Depression Scale; HADS, Hospital Anxiety and Depression Scale; KICA-Dep, Kimberley Indigenous Cognitive Assessment of Depression; LM, Logical Memory; MADRS, Montgomery-Asberg Depression Rating Scales; MCI, mild cognitive impairment; MMSE, Mini-Mental State Examination; MoCA, Montreal Cognitive Assessment; NH, nursing home; nRCT, nonrandomized controlled trial; PwD, patients with dementia; RCT, randomized controlled trial; STT-B, Shape Trail Test B; TAS-20, Toronto Alexithymia Scale–20; TMT-A/B, Trail Making Test-A/B; VBRT, Visual Benton Retention Test; VLT, Verbal Learning Test; VMSB, Visual Memory Span Backward; VMST, Visual Memory Span Test; WCST, Wisconsin Card Sorting Task; WHO-QOL-OLD, World Health Organization Quality of Life-OLD; WMS III, Wechsler Memory Scale III.
Characteristics of the other studies
Follow-up is in years. All studies were performed on elderly subjects in a community setting.
Chess
Rules
Chess originated in India in the VII century as chaturanga. Throughout history, the names of the pieces and their functions have evolved until they reached the current standardization, in which each player has one king, one queen, a pair each of bishops, knights and rooks, and eight pawns. The final objective of the game is to capture the opponent king. The game can also be won because the opponent has no time left. However, games can also end in a draw or stalemate when the player who has to move has no legal moves left and is not in check. Chess is a complicated game, with a rich theory for openings, midgame, and endgame, requiring elite players to memorize up to 40 moves per opening, with possible variants. It is also probably the most diffused board game, especially since the introduction of online chess, with sites such as chess.com allegedly having more than 100 million members at the time of writing.
Results
Despite its huge worldwide popularity, only two recent studies specifically investigated the effect of chess on cognitive impairment. Cibeira and colleague conducted a non-randomized controlled pilot study on 22 Spanish subjects in a gerontological complex composed of a daycare center and a nursing home (NH) [19]. Even though patients with moderate or severe dementia were excluded according to the protocol, more than half of the sample had severe cognitive impairment according to their Montreal Cognitive Assessment (MoCA) scores. The assignment to either 12 weeks chess training (two session per week) or control group was made according to the interest of participants themselves, but this did not result in significant differences at baseline except for gender (the control group comprised only women). No previous knowledge of chess was required. There was a significant group-by-time interaction only on MoCA, with significant improvements from baseline in the chess group, and significantly higher post-intervention scores in the chess group compared to controls (with a large effect size: Cohen’s d 0.84). There were no differences in Trail Making Test (TMT)-A or Visual Benton Retention Test. TMT-B was completed only by seven of the 22 participants in both assessments; in this sub-sample, performance significantly improved only in the chess group, without however significant differences between the two groups post-intervention. Chess did not seem to have any effect on depression over time. Indeed, Geriatric Depression Scale (GDS)-short form scores were lower in the chess groups at both time points, which may have influenced the willingness to participate in the intervention. Quality of life, measured with the World Health Organization Quality of Life (WHOQOL)-OLD scale, significantly improved only in the chess group compared to baseline scores, but no difference was observed between the two groups at either time point [19].
The study by Vale randomized 27 elderly Mexican women to either resistance training or resistance training plus weekly chess classes for 16 weeks. The authors found a significant improvement in Mini-Mental State Examination (MMSE) scores and quality of life measured with the WHOQOL-OLD scale in both groups (Cohen’s d = 1.0 for both outcomes). However, the chess group improved significantly more, with greater effect sizes [20].
The cross-sectional study by Deng investigated factors associated with prevalent dementia in community-dwelling Chinese elderly [21]. The authors evaluated 1781 community-dwelling elderly Chinese without a previous diagnosis of dementia. Prevalent dementia was defined by impairment on both MMSE, with different cut-offs based on education level, and instrumental activities of daily living. Chess and mahjong habits were grouped together, and coded as “every day”, “at least once a week”, “at least once a month”, “sometimes”, and “never”. Prevalent dementia was significantly higher, and almost double, in the group that never played compared to the group that played at least sometimes (11.74% versus 6.70%). Moreover, the frequency of subjects with prevalent dementia who played at least sometimes was significantly less compared to the one observed in the non-demented subjects (33/186, 17.7% versus 445/1,117, 39.8%, p < 0.0001) [21]. Given the cross-sectional nature of the study, it was not possible to exclude reverse causality.
Go
Rules
Go, or i-go in Japan, also known as wei-qi in China or baduk in Korea, may possibly be the most ancient board game still played nowadays, even though its origins are still controversial and lost in legends. While some historians claim that it was created during the third millennium BC, others retain that a more reliable estimate would be around the first millennium AD, when it was considered one of the four arts that a Chinese scholar should master. The game is played by two opponents on a 19x19 grid (versions with 13x13 or 9x9 cells are also popular), with some variability in how certain intersections are marked depending on the country. Each player has a number of stones, either black or white, that, contrary to chess, needs to be placed on the intersections of the grid, and not in the cells. The board is empty at the beginning of the game, and the black player starts. Go is a territorial game, and players place their stones in order to enclose intersections with them. Empty intersections surrounding a stone, or a group of stones, are called liberties. If a player manages to encapsulate a group of stones of the opponent with his own stones, so that the group has no liberties anymore, the stones are captured and removed from the board. A notable exception is when the group of stones of the opponent encloses at least two disconnected empty spaces, or eyes. In this case, the group of stones is said to be alive and cannot be captured. On the contrary, a group of stones which will inevitably be eventually enclosed is considered dead. Another exception is the case of seki, when an empty space is partially surrounded by both players in such a way that if either of them place a stone in that space, the opponent can capture the whole group with the next move. Another rule is that a player is not allowed to make a move that would create a repetition of the configuration before the opponent’s move. The purpose of the game is to gain as much territory (empty spaces) as possible. The game ends when both parties pass, or when there are no more moves available. The scoring is then made by counting the number of empty spaces enclosed by a player’s own stones, and subtracting the number of pieces captured by the opponent [1]. A system of adjunctive handicaps is usually used in order to balance forces, by giving white additional points to compensate for the fact that black begins to move. Although the game rules and principles are easier to understand than chess, Go seems to be extremely complicated, and further evidence of its difficulty is the delay (compared to chess) in the development of an artificial intelligence capable of consistently beating human masters. This goal was achieved only in very recent years, with the development of AlphaGo [22].
Results
Studies on Go have been performed by two groups in Japan and in China. No previous knowledge of Go was required in any of these trials. The group of Iizuka conducted two trials in Japan. The first trial randomized 22 subjects with moderate or milder dementia residing in NH to either 15 weeks of weekly Go lesson or a control group. The authors found a significant group-time interaction and a significant main effect of group for the total Digit Span Test (DST) and the backward task. The patients in the Go group significantly improved their scores, and these were significantly higher than those of the control group after the intervention (Cohen’s d for DST 0.85, and 0.87 for Digit Span Backward). The results for MoCA and digit span forward were not significant [23]. The second trial randomized 91 community-dwelling older adults to either face-to-face weekly Go classes, non-face-to-face weekly Go classes using a tablet or a control group undergoing a series of health education lectures. The authors found significant group-time interaction and significant main effects of group for the Visual Memory Span Test and the Visual Memory Span Backward, favoring both Go groups. In particular, scores in both groups improved after the intervention, and scores of the face-to-face group were significantly higher than those of controls at the end of the study. The effect size of the change in scores was greater between the face-to-face group and controls than between non-face-to-face group and controls. This was probably not driven by the degree of Go skills, as Go performance was comparable between the two intervention groups. No other significant differences were noticed in the other cognitive tests, including DST and TMT [24]. Since the face-to-face approach was used in all the other studies included in this review, only data for the former group were used for the meta-analysis.
Interestingly, the authors conducted an FDG-PET study on a subset of participants from the latter trial (13 from the intervention groups and 5 controls). They found significant increase in uptake in the left middle temporal gyrus and bilateral putamen in the Go group after the intervention, while FDG uptake increased significantly in the left superior frontal gyrus. There was a significant positive correlation between FDG uptake in the middle temporal gyrus and Go performance. In this subsample, the authors found a significant improvement in the Logical Memory II Test in both groups, and a significant worsening in the total DST in the control group [25].
Finally, Lin and colleague randomized 147 patients with a rather young-onset mild-to-moderate AD (mean age 42.2±12.2, mean MMSE 19.2±4.6) to either a long-time Go intervention (two hours daily), a short-time Go intervention (one hour daily), or a control group [26]. The study lasted 6 months. The intervention improved scores on the Montgomery and Asberg Depression Rating Scale, the Kimberley Indigenous Cognitive Assessment of Depression, and the Hospital Anxiety and Depression Scale, as well as the Global Assessment of Functioning scale and the RAND-36 for quality of life. There was no effect on the Toronto Alexithymia Scale-20. The authors also analyzed the levels of brain-derived neurotrophic factor (BDNF) and found that at baseline they were inversely correlated with the Clinical Dementia Rating (CDR), without differences between groups. However, at the end of the study serum BDNF was significantly higher in the Go groups compared to controls [26].
Mahjong
Rules
is unclear whether mahjong should be categorized as a board game or as a card game. It evolved as a variant of rummy, using tiles instead of cards. It is also substantially younger than chess and Go, being developed in China in the XIX century. A number of articles investigate the effect of mahjong on dementia by grouping it with card games [27–31]; in other cases, it has been grouped with chess [21, 33]. It is also not completely a game of skill, but a substantial influence of chance is present, which has been described as “gambling-like”. Contrary to other board games, it is generally played among four players. A mahjong set typically includes between 136 and 152 tiles, which are shuffled and placed faced down in two layers blocks. The tiles are grouped in three suits: dots, bamboo, and Chinese character, each containing four sets of number tiles (from 1 to 9). There are also two special types of tiles, namely honors (of two types: three sets of four dragon tiles, and four sets of four wind tiles) and bonus tiles (with two different sets of four tiles each for flowers and seasons). Players take turns throwing a dice which determines the number of additional tiles they receive, and then they arrange their own tiles. Then they draw one tile from the face-down pile, and discard a piece face up in the middle of the board; this discarded tile can be picked up and used by the other players irrespective of their turn. The players can then create pairs or other combinations of tiles, called melds. A pong and a kong are a set of three and four identical tiles, respectively, while a chow is formed by three suited tiles in a sequence. While players can pick up any tile from the discarded group in order to form a pong or a kong, to form a chow they can only use tiles discarded by the player on their left [34]. The meld formed using discarded tiles must be announced and showed to the other players; the players do not have to show their melds if they are formed by using tiles received at the beginning of each turn. Various combinations of tiles allow the player to win the game, which happens when a player creates a mahjong, formed by four melds and a pair. Each winning hand is scored according to a set of rules. A total of four hands are played, shifting the order of players by one, and the player with the highest score at the end of the four hands wins the game (for an illustrative example, one could refer to the following video by the Asia Pacific Mahjong association: https://www.youtube.com/watch?v=tRCb_LOkEmQ).
Results
All studies on mahjong included in this review were conducted in China. Four of the seven studies were performed on patients with dementia in NH, by the same group of Cheng and colleagues. Another study on patients with mild cognitive impairment or mild dementia in NH was performed by Zhang and colleagues, while two population studies investigated the role of mahjong in preventing cognitive impairment in the elderly.
The study by Zhang randomized 69 patients to either a 12-week intervention with mahjong three times per week or to a control group with no intervention. The authors found that the mahjong group significantly improved their scores in MoCA, functional activities of daily living (FAQ), and the Shape Trail Test B (STT-B, a culturally-adapted Chinese version of the TMT-B [35]), whereas no significant change was observed in the control condition. While the final scores for MoCA and STT-B did not differ between the two groups, FAQ were significantly lower in the game group at the end of the intervention compared to the control condition [34].
A series of four randomized controlled trials by Cheng investigated the effect of mahjong on patients with dementia in NH with various designs and outcomes. Two papers reported on different outcomes of the same studies, that randomized 110 patients to either mahjong, tai chi, or handicrafts for 12 weeks. Outcomes were collected up to 6 months after treatment. The studies showed that mahjong had a significant effect on MMSE, Digit Span Forward, Digit Forward Sequence (Cohen’s d 0.18 for MMSE and 0.16 for Digit Span Forward). For these measures and for CDR-sum of boxes (SB), there was also a significant group by time interaction, driven by a gradual decline in the controls performance, so that differences between mahjong and the control condition were progressively greater three and six months after intervention, when they reached statistical significance. There was no significant main effect or group per time interaction for Verbal Recall (immediate or delayed), Digit Span Backward or Digit Backward Sequence [36, 37]. Interestingly, only mahjong, and not tai chi, had a significant negative interaction with time on CDR-SB, meaning that patients randomized to mahjong tended to have lower scores over time compared to control subjects. However, there was no significant main effect of group on CDR-SB [37].
A previous study randomized 36 NH residents with dementia and at least moderate depressive symptoms to either mahjong, tai chi, or non-cognitively demanding handicrafts. The intervention lasted 12 weeks, and the outcomes were assessed at the end of the intervention period and after other three months of follow up. While the scores on the GDS remained substantially unchanged in the other two groups, the mahjong group significantly improved at the end of the intervention period (GDS 5.17±4.57 versus 8.42±2.50 at baseline) but returned to baseline after the three months in which the game had been discontinued (8.00±2.95 versus 8.42±2.50 at baseline). The authors found a significant group by time interaction on GDS, even after controlling for antidepressants, but no main effect for group [38].
Another trial by the same group randomized 62 patients to mahjong either twice of four times a week for 16 weeks, without a control condition. Outcomes were measured after the intervention ceased, and at one month follow-up. The authors showed that both groups significantly improved over time on MMSE, Digit Forward Sequence, and Verbal Learning Test, with similar patterns. Both groups also significantly improved on Digit Span Forward, especially in the group randomized to mahjong four times a week (in which the scores continued to increase even one month after the intervention). The effect of mahjong was larger on digit memory than on the other outcome measures (effect size baseline to follow-up ranged 1.33–1.56, while for MMSE it ranged 0.49–0.66) [39].
In all the aforementioned studies previous knowledge of the game was required, but it was required that the patients had not played for the past three or six months.
Two observational studies evaluated the effect of mahjong practice in elderly subjects without cognitive impairment. One of them grouped mahjong and chess together, and has already been described in the chess section [21]. In the other one, the authors evaluated data on 1,314 elderly Chinese in the Chinese Longitudinal Healthy Longevity Survey to address predictors of poor cognitive function and negative affect (including anxiety, depression, loneliness, fear, and unpleasant emotions). Cognitive function was measured with MMSE. They analyzed five waves of data over the course of 12 years, from 2002 to 2014, finding that playing mahjong had a positive and significant effect both on affective status (measured with a three-items Likert scale) and MMSE, in a model that included also gender, having chronic disease, marital status, co-residing with children, smoking, drinking, exercising and having social activities [40]. However, the study did not directly evaluate incident cognitive impairment as an outcome. Interestingly, around 20–25% of subjects in the study played mahjong, which makes it a very common leisure activity.
Ska
Rules
It was extremely difficult to gather information on the rules of ska. The paper by Panphunpho provides a rather unclear explanation [41], and attempts to find further information on the Internet were limited by the fact that all pages were in Thai. Nevertheless, after extensive searches involving several YouTube videos and the aid of Thai-speaking colleagues and friends (who were however not expert in board games), it emerged that the game of ska is probably not as diffused as the authors claim but limited to the north of Thailand. Moreover, there appear to exist several minor variations of the game, which seems to belong to the extremely broad family of mancala-type board games, in which a number of pieces have to be moved in a certain direction in a number of holes in a set of rows [1]. However, substantial differences from mancala exist. With those limitations in mind, the principles of ska can be summarized as follows: the board consists of two sides, each having 12 holes (named jooms) numbered from 1 to 12. Each player has 15 ska pieces, either black or white. A special piece, known as the city governor (tua kerd or chao muang), is placed upside down in the twelfth joom at the beginning of the game. This piece cannot be moved, unless another own piece moves across it; in this case, the tua kerd is flipped and can be used by the player. The other pieces will be initially placed in a middle area which is called elephant’s ear (hu chang). Players take turns to move their other ska pieces from their hu chang to their opponent’s side, counting from joom 1 to joom 12 of the opponent’s side, and then back to joom 7 to 12 of the player’s own side. In order to determine how many spaces players can move a piece, they take turns drawing cards or throwing dice with numbers from 1 to 6. For each turn, they can also draw cards or throw dice that show specific pictures that have specific functions, allowing players to move pieces in different ways. There seem to be a non-standardized variety of such pictures. The purpose of the game is to move all own pieces to own side of the board, completing a sort of “race” from the opponent’s side to one’s own. It is not clear whether the opponent’s pieces can be captured and how. What is clear, however, is that the game seems to be partially dependent on luck, and not only on strategy or memory.
Results
That said, the study by Panhpunpho randomized 40 non-demented community dwelling elderly (age 64.7±3.2) without previous experience of any cognitive demanding games to either 16 weeks of ska practice for 50 minutes three times per week or general health education activities [41]. At the end of the intervention period, the ska group significantly improved their scores in each subtest of the Wechsler Memory Scale III, in the TMT-A (with a very large effect size, Cohen’s d = 1.20), and in the Wisconsin Card Sorting Test compared to the control group. The authors also measured acetylcholinesterase activity before and after the study period, but did not find any significant difference between the ska group and controls [41].
Non-specified board games
In the study by Hughes, 942 American community-dwelling elderly subjects were followed from 1991 to 2002 to identify cases of incident dementia (defined as CDR of 1 or more) and investigate possible predictors or protective factors. Board games were evaluated separately from jigsaw puzzles, crosswords, and card games. Playing board games was coded in a dichotomous way. Only 3.5% of the participants played board games, a significantly lower percentage compared to other leisure activities (for instance, the prevalence of card games was 31.2%). Nevertheless, the proportion of participants who played board games was higher among those who did not develop dementia, although not significantly (3.73% versus 1.80%). While playing board games was not a significant protective factor in any multivariate model, the authors found that engaging in a greater number of activities and higher time commitment to leisure activities resulted in a lower risk of incident dementia [42].
Meta-analyses
We conducted meta-analyses for the change in different scales in WHO-QOL-OLD and cognitive outcomes from baseline to post-intervention, grouping together all board games. Raw data for all the tests in the paper by Cheng were kindly provided by the author upon request [36].
Regarding quality of life, the two studies used the same scale. We obtained distribution of the total score of WHO-QOL-OLD in the study by Vale by computing the mean and standard deviation of the sum of sub-scores provided in the paper. It emerged that chess significantly improved the scores of WHO-QOL-OLD compared to controls (I2 = 63%, MD 7.69, 95% CI 4.84–10.54, p < 0.00001). The forest plot of this meta-analysis is presented in Fig. 2.
Meta-analysis for quality of life.
For cognitive outcomes, a significant effect of board games was observed on MoCA (I2 = 75%, MD = 2.89, 95% CI 0.96–4.82, p = 0.003) and MMSE (I2 = 92%, MD = 2.61, 95% CI 0.45–4.76, p = 0.02). There was no difference in TMT-A change between board games groups and controls, although this was driven by the lack of effect seen in the study by Cibeira on chess [19]. There was no significant difference in TMT-B either, although the study by Zhang showed a significant difference on STT-B favoring mahjong [34]. Neither the Digit Span total score nor its sub-scores (forward and backward) showed significant differences, although there were significant differences favoring mahjong in all three scores. Finally, there was no significant difference between board games and controls in animal fluency. However, there was a significant difference between mahjong and controls favoring mahjong, and a significant difference between Go and controls favoring Go (even though both groups improved in the latter study). The control task in the Go study was a two-hour monthly lecture on health-related topic, while the control task in the mahjong study was simple handicrafts [24, 36]. The forest plots for these meta-analyses are shown in Figs. 2–5. The risk of bias for each interventional study is presented in Fig. 6.
Meta-analyses for MoCA and MMSE.
Meta-analyses for TMT.
Meta-analyses for Digit Span and Categorical Fluency (animals).
Risk of bias for randomized and nonrandomized controlled trials.
DISCUSSION
With the present meta-analyses, we were able to uncover evidence for several cognitive benefits of board games in elderly patients at risk of or suffering from dementia. Board games improve mental function, as measured by MoCA and MMSE, with a mean difference of 2-3 points after 3–4 months of intervention. Ska and Go, but not chess, seem to improve attention, as measured by TMT-A. Mahjong, but not Go or chess, may improve executive functions in this population. While Go does not improve the digit span total score, nor its sub-tasks, it seems to do so in the subgroup with dementia, and so seems to do mahjong.
Finally, mahjong seems to improve categorical fluency, while Go does not show this effect, possibly because of the specific features of the different games; indeed, while mahjong requires remembering several words or Chinese characters, Go pieces are just black and white stones. It could be interesting to evaluate whether the effect of mahjong on categorical fluency could be lost when played by subjects who do not speak Chinese, as in this case Chinese characters would be registered as pictures, and not as words. The effect of mahjong may also extend beyond the explored cognitive domains, possibly varying with age. Indeed, a study on middle-aged and older people showed that mahjong improved attention, reasoning, and short-term memory [43].
It has been showed that board games practice may have structural and functional biological effects. The study by Iizuka added to the biological effects of this game by showing that Go practice increased left middle temporal gyrus FDG uptake on PET, in parallel with an improvement in memory tasks [25]. Another study showed that long-term Go players had higher white matter connectivity in regions that are pivotal for the game, such as frontal, cingulum, and striato-thalamic areas related to executive function and working memory. No such effect was seen in premotor or parietal areas in long-term Go players [44], even though in studies with functional MRI this areas were also activated during Go (and with a right-predominant lateralization, contrary to chess, which shows a preferential left activity [3, 46]). However, there was no significant effect of Go on executive function in the included studies, possibly because the intensity and duration of the intervention were too short to sustain relevant biological effects in executive areas. Indeed, the study by Cheng showed additional positive effects with more frequent practice of mahjong, and it is possible that the same might apply to other board games [39]. Nevertheless, it is also possible that certain board games would not result in enhancement of specific functions per se. For instance, it has been showed that chess practice expertise did not result in better performances in executive tasks compared to matched controls, in line with what observed in our review [47]. This is, however, surprising, given the reported prefrontal activation in chess [48]. It may also be possible that the prefrontal activation required to play chess or Go remains limited to game-related tasks, without generalizing to other frontal-dependent abilities captured by standard neuropsychological tests.
The positive effect of board games on cognition may be also explained by biological effects, as the study by Lin showed an increased in BDNF with Go [26]. Similarly, another study showed an acute increase of BDNF in both elite and novice players after a session of chess, with more pronounced raises in the elite group [49]. This is analogous to what is observed for other interventions such as physical exercise and even acetylcholinesterase inhibitors, in which an increase in serum BDNF may mediate a protective effect against neurodegeneration [50]. Indeed, enhancement of endogenous BDNF or supplementation of exogenous BDNF are regarded as potentially interesting therapeutic strategies in AD [51].
The effects of board games may extend beyond cognition in the elderly population. For instance, chess may improve quality of life, while mahjong and Go improved depression in patients with dementia, even though the effect was short lived. Indeed, board games such as shogi (Japanese chess) have been used in cognitive-behavioral therapy programs to improve behavioral aspects such as negative thoughts [52]. It is, however, conceivable that much of the effect of board games could be explained by their social nature, as the face-to-face interaction seems to have additional positive effects on cognition [24]. This may be especially true for mahjong, which is played by four persons at once. Indeed, it is conceivable that a putative effect of board games in preventing depression/anxiety in the elderly may still have a potential role in preventing cognitive dysfunction, due to the intertwined relationship between psychiatric symptoms and incident dementia [8].
The cumulative evidence of short-term benefits of board games on cognition in individuals at risk of or with dementia from controlled trials supports the use-it-or-lose-it hypothesis [15]. It also makes the idea of reverse causality less likely, and it argues against an effect mediated by cognitive reserve, as education was balanced in all trials. However, board games practice may be itself a contributor of cognitive reserve, potentially protecting against the effect of advancing brain pathology [53]. Important to say, the two hypotheses are not mutually exclusive, and board game practice in the general population may be a marker of neuroselection, besides playing a neuroprotective role. Distinguishing between these two blending factors may be indeed tricky. Furthermore, in elderly subjects with significant board games mastery, it may be attractive to investigate the use specific tests, such as the Amsterdam Chess Test, to capture the first phases of cognitive decline in a more ecological way [54].
It is important to notice that many trials required no previous knowledge or at least no regular playing of the specific board game. While it is conceivable that in regular players an intervention with board games would not add much to what they already do (and there are also studies showing that even in experts chess ability naturally declines with age [55]), the fact that most of the participants with dementia could understand how to play and had a benefit from the intervention means that board games can be effectively taught to people with cognitive impairment, and may be a feasible intervention. However, the relatively high percentage of drop-outs in a number of studies imposes caution, as even low-intensity interventions may not be fully implementable in the fragile elderly with dementia.
Finally, the only study that did not show a positive effect of board games was the one by Hughes [42]. This study was also the only one looking at incident dementia with a considerable follow-up. Unfortunately, the type of board games played was not specified, and the percentage of players was nevertheless very low (only 3.5% of the sample). It is possible that the kind of games played by the subjects did not include only traditional board games, but also less cognitively demanding games. Indeed, the other observational study with a comparable follow-up showed that mahjong was protective against cognitive deterioration on MMSE, even though the authors did not look at incident dementia [40]. In this latter study, the proportion of subjects playing board games was also significantly higher than the one in the paper by Hughes, possibly reflecting cultural differences between the US and China.
Our study has both strengths and limitations. Contrary to existing reviews on games and cognition in general [56], we included only studies focusing on preventing or slowing down cognitive decline in populations at risk or already suffering from cognitive decline. Moreover, we specifically excluded card games, as we hypothesize that they may entail different neural substrate and require a separate analysis. Finally, we gathered data from various resources available on the topic, including also non-randomized and observational studies, and included more cognitive outcomes in comparison with other reviews on games and cognition in the elderly (such as the one ongoing here: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=296857).
Our study has also some limitations. Most of the included studies are pilot trials, with a small number of participants and a significant number of drop-outs in some cases. While the duration of the interventions is similar across studies, different selection criteria and protocols increase the heterogeneity of the data and limit the generalizability of our findings. Also, no trial performed a head-to-head comparison of board games, therefore we could not conclude if any of them could be generally better for cognition, or whether more games could have a synergistic or additive effect. Given these aforementioned limitations, it was not always possible to provide conclusive answers on the effect of board games on cognition. However, we believe that our data could represent a promising starting point for further research on the topic.
Conclusions
Traditional board games such as chess, Go, or mahjong may represent a feasible cognitive intervention to prevent dementia and slow cognitive decline. They can also improve quality of life and depression in elderly patients at risk of or suffering from dementia. Different games seem to differentially affect specific cognitive domains, such as attention, executive functions, or memory. The effect on cognition may be mediated by functional and biological factors, such as an increase in BDNF or regional brain activity. While more research is needed to confirm and extend the results of interventional trials in larger samples, board games practice should be incentivized, especially in the elderly.
Footnotes
ACKNOWLEDGMENTS
The authors have no acknowledgments to report.
FUNDING
This publication was produced with the co-funding European Union –Next Generation EU, in the context of The National Recovery and Resilience Plan, Investment Partenariato Esteso PE8 “Conseguenze e sfide dell’invecchiamento”, Project Age-It (Ageing Well in an Ageing Society).
Italian MUR Dipartimenti di Eccellenza 2023–2027 (l. 232/2016, art. 1, commi 314–337).
CONFLICT OF INTEREST
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
DATA AVAILABILITY
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.
