Pilot Randomized Controlled Trial of the GO Game Intervention on Cognitive Function

Abstract

This study investigated the effects of an intervention using the game “GO” on cognitive function in nursing home residents and evaluated the acquisition of GO according to each stage of dementia. Participants were randomly assigned to either the GO intervention group or a control group, and the intervention was performed once weekly for 15 weeks. Cognitive tests were conducted before and after intervention, and 17 participants were included in the final analysis. Analysis of covariance demonstrated that in the intervention group, the digit span total score significantly improved and the digit span backward score was maintained, whereas these scores decreased in the control group. All participants, including those who had moderate dementia, acquired the rules of the game, and participants with mild cognitive impairment and mild dementia could play the game successfully. This study indicates that GO might improve the cognitive function of residents living in nursing homes.

Introduction

Dementia is a major topic of concern for aging societies. Recently, the dementia population has reached over 46 000 000 worldwide. ¹ Thus, measures to prevent or mitigate dementia are urgently required. It has been pointed out that the cognitive functions of older adults living in nursing home facilities tend to decline faster than physically and socially active people. ² To solve this problem, several intervention programs utilizing intellectual activities have been introduced for facility residents with low physical and cognitive function. Among them, cognitive leisure activities are garnering interest in researchers.

Several observational studies have revealed the effects of cognitive leisure activities on cognitive function. Stern and Munn defined cognitive leisure activities as those requiring a mental response from the individuals taking part in the activities (eg, reading). ³ They conducted a systematic review of the relationship between cognitive leisure activities and cognitive function and reported that cognitive leisure activities may be beneficial in preventing the risk of dementia. ³ Yates and colleagues also conducted a systematic review and meta-analysis, and their results showed that the higher the frequency of participating in cognitive leisure activities, the lower the risk of developing dementia. ⁴ More specifically, among all cognitive leisure activities, a previous study showed that reading, playing board games, playing a musical instrument, and dancing contribute to reducing the risk of developing dementia. ⁵ Furthermore, in a French cohort study, it was shown that playing board games may have possible beneficial effects on reducing the risk of dementia, cognitive decline, and depression. ⁶

Intervention studies utilizing cognitive leisure activities such as board games have also been conducted. Board games are expected to be useful in facilities since they involve many aspects of intellectual stimulation and even people with low physical function can enjoy them. In particular, there are several intervention studies utilizing mahjong. ^{7 –9} Mahjong is a famous board game in Asian countries, usually played together by 4 people. Mahjong set includes 136 to 152 tiles and the rules for winning and calculating scores are rather complex. ⁷ Interventions utilizing mahjong have shown positive intervention effects on cognitive function across a range of cognitive domains including attention, short-term memory, and working memory, and these effects were long-lasting. ⁸ Furthermore, mahjong was shown to delay the progression of dementia. ⁹ However, the rules of mahjong are complex and difficult for those with low cognitive function to learn. Since high cognitive function is required for beginners to learn the game, previous studies on mahjong conducted by Chen targeted only those who already knew how to play the game. ^{7 –9} Therefore, it is preferable to select an intervention that allows everyone who wants to participate to join regardless of their level of cognitive function.

Besides mahjong, there are other programs for people with low cognitive function, such as music therapy, ¹⁰ art, ¹¹ and learning therapy that utilize arithmetic and reading aloud. ¹² However, intervention studies in which older adults learn new things are limited to the healthy population. For this reason, it is not clear whether older people with low cognitive function can learn new things or how difficult it is for them to master certain skills. There is an indication that learning new things is effective for preventing cognitive decline, ¹³ and it is possible that this concept can be applied not only to healthy older adults but also to those with low cognitive function. Therefore, this topic is worth investigating.

In this study, we focused on the game “GO,” which consists of simple rules and provides high intellectual stimulation. GO is a famous board game in Asian countries, particularly Japan, China, and Korea, and it is gaining popularity in the United States and Europe. GO has only 3 main rules, which are very simple for beginners to memorize. Although experienced players are better than beginners, the difficulty level can be adjusted according to the cognitive function of the player. This was observed in an intervention study using GO for people with early-onset Alzheimer’s disease. ¹⁴ In that study, behavioral and psychological symptoms of dementia were reduced and brain-derived neurotrophic factor levels were improved by learning to play GO; however, there was no intervention effect on global cognitive function as measured by the Mini-Mental State Examination (MMSE). ¹⁵ It is important to mention that the MMSE, an examination of global cognitive function, was the only cognitive test used in that study. Since cognitive function is composed of multiple domains, it is important to conduct cognitive tests for each domain in addition to global cognitive function.

When playing GO, it is necessary to plan the next move based on the previous and current game situation, and players must pay attention to many areas of the wide game board. For these reasons, GO is considered effective for maintaining and improving attention and working memory. A previous study found positive effects of GO on cognitive function in children with attention deficit hyperactivity disorder by means of the digit span test, ¹⁶ a test that evaluates attention and working memory. However, intervention studies with older adults have not yet been conducted, and it is important to reveal the effects of GO in that population.

Therefore, the purpose of this study was to conduct a GO game intervention for older adults in order to investigate its effects on their cognitive function. We also investigated differences in the acquisition of the game GO according to the level of cognitive function.

Methods

Study Design and Participants

A randomized controlled open-label trial was conducted at 2 nursing homes in Kanagawa, Japan. In the first facility, the intervention was conducted from February to October 2015, and in the second facility from August 2016 to April 2017. Originally, we planned to start the intervention at the same time in both facilities; however, it was delayed at 1 facility due to a facility-related issue. Inclusion criteria were 65 years of age or older, Clinical Dementia Rating (CDR) scale ¹⁷ of 2 or less, and no previous experience playing the game GO. All etiologies of cognitive decline were included. Participants were excluded if they were receiving acute treatment for physical or psychological diseases, had a CDR of 3, or were bedridden. Clinical Dementia Rating was assessed by a medical doctor. Figure 1 shows a diagram of this study based on the Consolidated Standards of Reporting Trials (CONSORT) statement. ¹⁸ Nursing home residents and their families received written and oral information about this study, and written informed consent was obtained from 40 residents who agreed to participate in this study. After receiving informed consent, participants were selected based on the inclusion criteria. Five participants were excluded because of severe dementia (CDR of 3) and 2 were excluded due to acute health problems. The remaining 33 participants were randomly assigned to either the intervention group, which attended GO classes, or to the control group, which received the usual care. Participants were allocated to either the intervention or the control group at the same ratio in both facilities.

Figure 1.

Consolidated Standards of Reporting Trials flow diagram.

Ethical Approval

Ethical approval was obtained from the institutional review board and ethics committee of the Tokyo Metropolitan Institute of Gerontology (Acceptance No. 84, 12, 2014).

Intervention

A professional GO game player belonging to the Nihon Ki-in and 3 GO instructors visited the 2 nursing home facilities to hold the program once a week for 1 hour, for a total of 15 classes. Each 1-hour session consisted of a lecture on the basic rules and techniques of the game GO (15 minutes), solving GO game exercises (15 minutes), and playing games (30 minutes). The interventions conducted at both facilities consisted of the same program. All participants received the same lectures regardless of their level of cognitive function.

The instructors created original GO game exercises for this program. The instructors presented all exercises using a GO game board and GO stones and no exercises were presented in writing. The instructor adjusted the difficulty level and the number of questions according to participants’ level of cognitive function and speed of solving the exercises. The easiest task was learning the following 3 basic rules of GO. (1) One player uses the black stones and the opponent uses the white ones. The players take turns placing the stones on the intersections of the lines. (2) When one of the stones is surrounded by the opponent’s stones, the surrounded stone is captured by the opponent and removed from the board. (3) A player wins when they fully surround a larger total area of the board than their opponent. If the instructors judged that a participant had a good understanding of the basic rules and would enjoy learning more advanced techniques, they taught advanced GO game techniques to those participant (defined techniques of GO).

Participants who learned advanced GO techniques played the game against instructors or other participants who had achieved the same level. Participants who needed a lot of support to learn the basic rules conducted tasks using only some of the basic rules, so that they could easily play simple game situations and enjoy playing the game. In addition, they practiced a traditional method of playing GO called “Kifu-narabe,” which consists of learning a model game.

Outcome Measures

Cognitive function

Cognitive tests were conducted to assess the effects of the GO game training. We set attention and working memory as the main outcome measures and global function as the secondary outcome measure. Assessments were carried out just before the start of the program and after completion of the 15th GO class.

Global cognition

The baseline characteristics of the participants were assessed using the MMSE. The maximum score on the MMSE is 30, and the cutoff score to screen for dementia is 23 / 24. The Japanese version of the Montreal Cognitive Assessment (MoCA-J) ¹⁹ was used to assess changes in global function before and after the program. The MoCA-J is a brief cognitive screening tool for mild cognitive impairment (MCI) with a maximum score of 30 and a clinical cutoff score of 25 / 26.

Attention, working memory, short-term memory

The digit span test, which is included in the Wechsler Adult Intelligence Scale version III, ²⁰ was used to evaluate attention. In addition to attention, the digit span test also contains tasks to assess short-term memory and working memory.

Digit Span forward task

This was used to measure short-term memory in addition to attention. The examiner reads aloud a random number sequence and the participants are asked to repeat it.

Digit Span backward task

This was used to measure working memory and complex attention. The examiner’s method is the same as the forward task, but participants are asked to repeat the sequence in reverse.

Staging of dementia

We evaluated the progression of dementia in the participants using the CDR. The CDR is an assessment tool for staging the severity of dementia. It consists of 6 cognitive categories (memory, orientation, judgment and problem solving, community affairs, home and hobbies, and personal care). In this study, a medical doctor scored each category by interviewing the participants and their caregivers and then judged the total CDR score. Participants were assigned a rating of healthy (CDR of 0), mild dementia (CDR of 1), moderate dementia (CDR of 2), or severe dementia (CDR of 3). A CDR of 0.5 was regarded as possible dementia/MCI.

Other assessments

In addition to cognitive function, we obtained basic information such as age, gender, education level, medical history, internal medical history, and activities of daily living. Activities of daily living were evaluated using the Barthel index. ²¹

GO game performance

Three GO game instructors who did not engage in analysis of the results evaluated the GO game performance of the participants during the program. GO game performance assessed whether the participants had learned the basic rules of GO or the more advanced defined techniques of GO and the results of playing games. Whether or not the participants asked for the instructor’s support when practicing the GO game exercises and playing games was also taken into consideration. Each item was evaluated using symbols. If the participants could work through the exercises without the instructor’s support, they were given a “+”; if the participant always received a lot of one-to-one support from the instructor and still found it difficult to work through the exercise, they were given a “−”; and if the participant could work through the exercises with a little advice from the instructor, they were given a “±.” When participants could play against other advanced players or instructors, they were considered to have fully learned how to play the game (playing games was given “±” or “+” based on whether instructor support was received or not, respectively, the same as for the GO game exercises.). If the participant could not play games, even with the instructor’s support, it was deemed impossible for them to completely master GO (playing games was given “−”).

Statistical Analysis

All analyses were performed with SPSS version 23. Baseline characteristics between the intervention and control groups were compared using independent sample t tests, and comparisons for gender were conducted using the χ² test. Differences in cognitive test scores between groups over time were analyzed using a series of mixed model analysis of covariance (ANCOVA). Age and education level were set as covariates. A P value below .05 was considered statistically significant.

Results

Characteristics of Participants

Table 1 shows the characteristics of all participants. The results of independent sample t tests and χ² tests showed no significant differences in age, gender, education level, MMSE score, and MoCA-J score between the intervention and control groups.

Table 1.

Baseline Characteristics of All Participants.

Characteristic	Intervention Group (n = 9)	Control Group (n = 8)	P Value
Age, mean (SD), years	89.1 (4.1)	89.1 (6.6)	.996
Gender (female/male), N	8/1	7/1	1.000
Education, mean (SD, years	11.3 (2.7)	11.6 (2.5)	.824
Barthel index, mean (SD), score (0-100)	62.2 (25.5)	64.3 (28.7)	.872
CDR (0/0.5/1/2), score (N)	0/2/2/5	0/1/2/5
MMSE, mean (SD), score (0-30)	20.0 (4.8)	17.0 (4.6)	.212
MoCA-J, mean (SD), score (0-30)	13.6 (5.6)	10.6 (5.6)	.286

Abbreviations: CDR, Clinical Dementia Rating; MMSE, Mini-Mental State Examination; MoCA-J, Japanese version of the Montreal Cognitive Assessment; SD, standard deviation.

Compliance With the GO Game Program

Eight participants in the intervention group dropped out of the program for the following reasons: (1) 1 refused to perform the cognitive tests at baseline, (2) 2 did not attend any of the GO classes, (3) 2 came to the first class but did not attend thereafter, (4) 2 were hospitalized with severe medical disorders, and (5) 1 died. The intervention group had 70% attendance over the 15 classes. Of the participants assigned to the control group, 6 participants left the study for the following reasons: (1) 2 refused to perform cognitive tests at baseline, (2) 2 refused to conduct follow-up tests, and (3) 2 were hospitalized with severe medical disorders. Two participants who were originally assigned to the control group strongly refused to participate in the control group, so these 2 participants were also excluded from the analysis. Therefore, 9 participants in the intervention group and 8 in the control group were included in the final analyses.

Effect on Cognitive Function

Table 2 shows the cognitive test scores before and after intervention, the percentage of participants with improved scores after intervention, and the ANCOVA results. In the digit span test, the main outcome measure in this study, there were significant interactions between groups and times for the total test score and the backward task. Furthermore, simple main effects for the total test score in the intervention group were significantly higher compared with the control group after intervention (P < .05), and the results of the intervention group significantly improved after the intervention (P < .05). The digit span backward task also showed simple main effects in the control group. The scores of the control group were significantly lower (P < .05) than the intervention group after intervention and significantly decreased after the intervention (P < .05), whereas no significant change was observed in the intervention group. Regarding global cognitive function, the interaction between group and time for the MoCA-J scores was marginally significant (P < .1). The simple main effect was also marginally significant with lower scores in the control group compared with the intervention group (P < .1).

Table 2.

Cognitive Test Scores and ANCOVA Results Before and After Intervention.

	Intervention Group (n = 9)			Control Group (n = 8)			ANCOVAa
	Before	After	Participants With Improved Scores (%)b	Before	After	Participants With Improved Scores (%)b	Interaction		Main Effect
							Time × Group		Time		Group
							F	P Value	F	P Value	F	P Value
MoCA-J score	13.6 (5.6)	14.4 (7.0)	55.5	10.6 (5.6)	9.1 (5.9)	12.5	3.835	.072	1.015	.332	2.636	.128
DST total score	11.2 (4.5)	13.1 (3.7)	66.6	10.8 (3.3)	9.3 (2.9)	12.5	7.900	.015	0.448	.515	1.993	.182
DSFT score	7.6 (4.0)	9.2 (2.4)	66.6	7.5 (2.1)	7.3 (2.0)	37.5	2.341	.150	0.141	.714	0.914	.356
DSBT score	3.5 (2.0)	3.8 (1.7)	33.3	3.3 (1.6)	2.0 (1.0)	12.5	5.493	.036	2.885	.113	2.212	.161

Abbreviations: ANCOVA, analysis of covariance; DSBT, digit span backward task (score range 0-14); DSFT, digit span forward task (score range 0-16); DST, digit span test (score range 0-30); MoCA-J, Japanese version of the Montreal Cognitive Assessment (score range 0-30). ^aAge and education level were set as covariates, and the main effects were not statically significant.

^bIt shows the percentage of participants with improved scores.

GO Performance

Table 3 shows the baseline characteristics of participants in the intervention group and individual status of acquisition of the game GO at the end of the program. All participants, including those with moderate dementia, understood the basic rules. In addition, all participants with mild dementia and MCI and 3 of 5 participants with moderate dementia could solve advanced level GO game exercises using defined techniques of the game. Participants with MCI and mild dementia and 1 of 5 participants with moderate dementia were able to play GO. The mean age of the participants who mastered GO was 90.0 (standard deviation [SD] = 4.7]) years, the mean MMSE score was 23.4 (SD = 3.3), and the mean MoCA-J score was 17.6 (SD = 3.3). The average age of participants who could not master GO was 88.0 (SD = 3.7) years, their mean MMSE score was 15.7 (SD = 1.8), and their mean MoCA-J score was 8.7 (SD = 3.4). Independent sample t tests on the participants’ baseline characteristics showed no significant difference in age between those who mastered GO and those who did not. However, MMSE scores (P < .01) and MoCA-J scores (P < .01) were significantly lower among those who could not master GO.

Table 3.

Individual Status of Acquisition of the Game GO in the Intervention Group.

Participants	Characteristics					GO Game Performancea
	Age (years)	Gender	CDR	MMSE	MoCA-J	Basic Rules	Advanced Techniques	Playing Games
A	89	F	0.5	27	22	+	+	+
B	83	F	0.5	27	20	+	+	+
C	90	F	1	22	17	+	+	+
D	96	F	2	21	15	+	±	−
E	92	M	1	20	14	+	±	±
F	88	F	2	17	13	+	±	±
G	93	F	2	17	10	+	±	−
H	87	F	2	16	6	±	−	−
I	84	F	2	13	6	±	−	−

Abbreviations: CDR, Clinical Dementia Rating; F, female; M, male; MMSE, Mini-Mental State Examination; MoCA-J, Japanese version of the Montreal Cognitive Assessment. ^a+ indicates the participant could work through the exercises without the instuctor’s support.− indicates the participant received a lot of support from the instructor and still found it difficult to work. ± indicates the participants could work through the exercises with a little advice from the instructor.

Discussion

The aim of this study was to reveal the effects of playing the game GO in older adults with cognitive decline and to investigate differences in the acquisition of the game according to the level of cognitive function. We found that the intervention group showed improved attention and working memory scores, while the control group showed declines in these scores. In addition, the assessment of GO performance after the intervention showed that even older adults with cognitive decline could understand the rules of GO and both patients with MCI and mild dementia were able to completely master the basic techniques of the game. Furthermore, it was possible for all participants to complete the program by adjusting the level of difficulty of the exercises. This suggests that even people with cognitive decline can learn and acquire new techniques, and the game GO was considered an effective and versatile tool that can be applied to people with a wide range of cognitive functions.

In this study, participants in the intervention group who could master GO showed significantly improved digit span test scores and thus improved attention. Further analysis of the digit span backward task indicated that working memory was significantly reduced in the control group and maintained in the intervention group. Several studies have examined the effect of the game GO on brain function. Chen and colleagues examined the brain activity of university students when looking at images of GO by using functional magnetic resonance imaging. Results showed that when looking at images of the game, blood flow increased in the dorsal side of the prefrontal cortex, parietal lobe, occipital lobe, posterior temporal lobe, and the primary somatosensory cortex as compared to when looking at a meaningless array of stones. ²² Based on this research, Kim and colleagues hypothesized that playing GO must involve attention and executive function, and therefore, they conducted an intervention study on 17 children with attention deficit hyperactivity disorder. ¹⁶ Their results of the digit span test, digit span forward task, and Children’s Color Trial Test showed significant improvement in the intervention group compared to healthy children in the control group. Our study confirms the findings of previous studies in that it also shows improvements in attention and working memory.

To the best of our knowledge, this is the first study showing that older adults with cognitive decline can learn a new technique along with an evaluation of their acquired learning of GO techniques. Regardless of the learning situation, it was shown that cognitive function could be maintained or improved by continuing to participate in the program.

The findings obtained in this study suggest that learning a new technique is effective in suppressing cognitive decline and provide evidence to show that this concept can be applied to cognitively healthy older persons and those with cognitive decline. In addition, our findings suggest that the act of learning itself may be effective in mitigating cognitive function even if the technique is not completed mastered.

GO is a learning tool that can be applied to people with various cognitive functions for 2 reasons. First, by adjusting the level of difficulty of the present program, it was possible for participants with a wide range of cognitive functions to continue participating in the program. Second, all participants, including patients with moderate dementia, were able to understand the basic rules of the game. Furthermore, it was possible for those with MCI to play GO without the support of instructors. Three participants in this study voluntarily continued to play the game after completion of the program, such as by working on a collection of GO game exercise books they purchased themselves. Thus, in addition to being a game that even older adults with cognitive decline can play, GO can be enjoyed even if there is no opponent. Therefore, the game GO can be easily utilized as part of a program to prevent cognitive decline at nursing home facilities.

The results of the digit span backward task showed the working memory of the control group significantly decreased in a short period of 5 months. Previous studies have also shown that the cognitive function of facility residents significantly decreased in a short period of several weeks to half a year. ^23,24 Older adults who are in a facility have few external stimuli and spend most of their daily life sitting or in bed. This means there are few opportunities to interact with other people. For these reasons, cognitive function is likely to rapidly decline. Furthermore, it has been pointed out that facility staff often underestimate the capabilities of residents, and there are few opportunities for these older adults to show their inherent abilities. ²⁵ This may be a vicious circle for facility residents; therefore, it is meaningful not only to improve but also to maintain the cognitive function of facility residents. It is also important to positively adopt programs that have such effects. The program in the present study may be a suitable option for adding intellectual stimuli from such aspects.

This study had a few limitations that should be acknowledged. First, it is difficult to conclude that the results obtained were due to the effects of learning the game GO itself. This is because other factors may have been associated with implementation of the program; specifically, the increased interaction with staff or other participants may have influenced the results. Previous studies have shown that greater interaction with others results both in greater cognitive function and a lower incidence of dementia. ^26,27 Therefore, it is expected that increased social interaction for facility residents who do not often experience contact with others will provide strong cognitive stimulation. During the intervention program, it was difficult to completely exclude elements of interpersonal contact; however, this point has been a long-standing problem in intervention studies targeting cognitive function and future examination is required. Second, many participants dropped out due to reasons such as acute health problems. Because most residents in nursing homes are near 90 years of age and frail, dropping out due to changes in physical or mental state is inevitable. Furthermore, the sample size was small, so a large sample size is needed to clarify this intervention effect, and for that purpose, it is considered appropriate to conduct clustered Randomized Controlled Trial in the future.

Conclusions

Older adults with cognitive decline living in nursing homes showed improved or maintained attention and working memory by regularly playing the game GO. Moreover, it was possible for all participants to learn how to play GO, even those with dementia, thereby demonstrating the possibility of using GO to improve and maintain cognitive function. Also, it is possible for nursing home facility residents to continue participating in a GO game program by adjusting the level of difficulty of the tasks. Therefore, GO seems to be an appropriate tool for people with a wide range of cognitive functions.