Effectiveness of Functional Mobile Games to Enhance Attention in Individuals with Mild Cognitive Impairment: A Study Using The Montreal Cognitive Assessment

Abstract

Purpose

This study aimed to evaluate the effects of functional games on attention in older adults with mild cognitive impairment (MCI), addressing the limited focus on specific cognitive domains in existing research. Additionally, it explored how functional games may stimulate multiple cognitive domains, contributing to overall cognitive improvement.

Materials and Methods

This randomized controlled trial recruited 80 participants with MCI from April to June 2023 through outpatient visits at the Neurology Department of Incheon St. Mary's Hospital, Catholic Kwandong University, and public advertisements. Participants were randomly assigned to the treatment or control group. The treatment group engaged in functional games at least three times per week for 12 weeks. Measurements were conducted at baseline and after the intervention period.

Results

Among 71 participants, improvement in the attention domain of the Montreal cognitive assessment (MoCA) was more significant in the treatment group (42.11%) than in the control group (18.18%, p=0.003). Clinical dementia rating (CDR) analysis showed a slower cognitive decline in the treatment group (p=0.033), while no significant changes were observed in mini-mental state examination scores.

Conclusion

The intervention demonstrated a significant improvement in attention as measured by MoCA and a slower progression of cognitive decline as indicated by CDR, highlighting its potential effectiveness in enhancing cognitive function and mitigating cognitive decline in older adults. These findings support using such interventions as a viable strategy for cognitive improvement.

Graphical Abstract

INTRODUCTION

Preserving and improving the cognitive functioning of older adults has been an important public health issue, and researchers have been designing and implementing a variety of programs and activities to prevent the risk of dementia by increasing the levels of cognitive function of older adults.¹ Research has provided substantial evidence that the design and application of games can have a positive effect on cognitive processes such as attention and executive functions.¹,² A recent longitudinal study has provided evidence that cognitively stimulating activities such as games and puzzles are associated with higher levels of memory, working memory, and attention and processing speed among older adults with mild cognitive impairment (MCI).³

In light of the importance of cognitively stimulating activities, cognitive training is considered an effective approach that adopts structured guidance on specific cognition-related tasks to preserve or enhance specific cognitive abilities such as memory, attention, concentration, and problem solving through various training methods.⁴ Research has demonstrated that various types of cognitive training activities influence neuroplasticity, enhance mental capacity, and increase interest in improving cognitive function through cognitive training.⁵,⁶

Traditional cognitive training interventions are designed as face-to-face interactions in clinical settings that require significant financial, time, and space resources, as well as monitoring of participant adherence.⁷,⁸ To minimize these challenges that are associated with traditional cognitive training interventions, research suggests the use of individualized, customized digital health applications for older adults. Recently, game-driven cognitive enhancement programs using mobile technology have leveraged the portability and accessibility of mobile applications to deliver engaging and effective cognitive training.⁹,¹⁰ This is facilitated by engaging user interfaces and the ability to incorporate user feedback in real time, which allows for efficient content usage and difficulty adjustment in online environments.

Substantial evidence supports the efficacy of game-driven cognitive enhancement programs using mobile technology to improve cognitive function in older adults.⁹,¹⁰ For example, Nouchi, et al.⁷ investigated the effects of the “Brain Age” brain training game using the Nintendo Wii on older adults without cognitive impairment. They found that older adults who participated in this training game for 4 weeks (more than five times per week) showed improvements in executive function and processing speed that demonstrated the potential for cognitive enhancement even with short-term training. In addition, some clinical studies demonstrating the impact of specific functional games on executive function, memory, and attention reported significant improvements in cognitive function scores among older adults with MCI.¹,¹¹ These studies highlighted that game-based training for older adults with MCI and early-stage dementia is associated with improvements in mini-mental state examination (MMSE) and Montreal cognitive assessment (MoCA) scores, indicating the potential for cognitive function enhancement through engaging and intuitive games.¹²

While research has reported that cognitive training game programs using mobile technology have positive effects on cognitive function improvement,⁹,⁷,¹² these previous studies have generally reported improvements in overall cognitive function, though few studies have focused on evaluating the effects of cognitive training game programs using mobile technology on specific cognitive domains. In addition, there is a lack of clinical trials investigating the impact of mobile-based, game-driven cognitive training programs among older adults with MCI in an international context. Therefore, the primary objective of this study was to assess the effectiveness of a mobile-based, game-driven intervention for improving attention in older adults with MCI. Additionally, considering that functional games may stimulate multiple cognitive domains, the study also evaluated the effect on overall cognitive functioning through the use of these functional using the clinical dementia rating (CDR).

MATERIALS AND METHODS

Participants

Participants were recruited from the Neurology Department of Incheon St. Mary's Hospital, Catholic Kwandong University, and through public advertisements. After explaining the purpose, methods, and observational parameters of the study, participants who provided written informed consent were randomly assigned to either the experimental or control group. The experimental group consisted of individuals receiving pharmacological treatment for dementia who were instructed to use functional games for 12 weeks. The control group included individuals receiving pharmacological treatment only. The study inclusion criteria included: 1) individuals aged 65 years or older who were diagnosed with MCI or mild dementia and were currently undergoing pharmacological treatment, 2) diagnostic criteria corresponding to an MMSE score of 16–26, 3) standard pharmacological treatment with Acetylcholineesterase inhibitors, such as Donepezil, Galantamine, or Rivastigmine for more than 3 months (to ensure stability and sufficient pharmacological effect), 4) individuals who understood the study procedure and agreed to participate, and 5) individuals capable of understanding game instructions and operating the game device. Exclusion criteria included: 1) individuals who failed to complete written informed consent prior to screening, 2) individuals unable to consistently participate due to physical, cognitive, or technological limitations (e.g., difficulty using mobile devices), and 3) individuals who had previously participated in other gaming or similar cognitive training programs prior to enrollment. We declare that all procedures performed in this study involving human participants were approved by the Institutional Review Board of Incheon St. Mary's Hospital, Catholic Kwandong University (IRB No. IS23EISE0031). Informed consent was obtained from all individual participants included in the study.

Sample size calculation

Sample size estimation was initially performed using G*Power 3.1 software.¹³ A medium effect size of Cohen's d=0.50, based on findings from previous meta-analyses of cognitive training interventions for older adults,⁸,¹⁴,¹⁵ was assumed. With a significance level of α=0.05 and power (1−β)=0.80, the estimated required sample size was 64 participants in total (32 per group), using a two-tailed independent t-test with equal allocation. Accordingly, 30 people with statistical significance were targeted for each group, and 40 people were recruited for each group to account for a 20% dropout rate.

However, the primary analysis of this study was conducted using a linear mixed-effects model, which generally requires a larger sample size to achieve equivalent statistical power. Recalculating based on a repeated-measures design with the same effect size (converted to Cohen's f≈0.25), the required sample size increases to approximately 52 participants per group. Although this final sample size may be slightly underpowered under mixed-model assumptions, it was determined to be ethically and practically feasible within the constraints of the clinical research setting.

Study design

This study was conducted as a randomized controlled trial (RCT) involving older individuals with MCI. This RCT was reviewed and approved by the Institutional Review Board (IRB) of [Incheon St. Mary's Hospital, Catholic Kwandong University] prior to participant enrollment. The study was conducted in accordance with the Declaration of Helsinki. Although the protocol was not registered in a public trial registry such as ClinicalTrials.gov, a complete study protocol and statistical analysis plan (SAP) were established before recruitment and are available upon request. Participants were recruited through public advertisements and outpatient visits at the Neurology Department of Incheon St. Mary's Hospital, Catholic Kwandong University. The research team explained the study's purpose, methods, and observational parameters in detail, and participants provided informed consent before screening based on the study inclusion and exclusion criteria. Eligible participants were randomly assigned to the experimental group (n=40) or the control group (n=40). During randomization, the block randomization method was used to assign participants to the experimental and control groups in the order of “ABAB AABB BBAA BABA ABBA BAAB BABA BBAA AABB ABAB.” All participants attended two assessment visits. The first visit (Screening & Baseline) involved collecting MMSE, MoCA, and CDR scores, administered by certified researchers separate from the researcher conducting the study. The second visit took place after the 12-week intervention period in which reassessments were conducted. The intervention group engaged in functional game training at least three times per week for 12 weeks, during which game usage was monitored to ensure adherence. To minimize potential measurement bias, outcome assessments (MMSE, MoCA, and CDR) were conducted by trained clinical raters who were blinded to group assignment. Clinical tests were evaluated by experts who obtained certification for the test. These assessors were not involved in the recruitment or intervention delivery process and conducted the cognitive evaluations in a separate clinical setting from the intervention sessions. Blinding procedures were strictly maintained to reduce the possibility of observer bias, especially for CDR, which involves subjective clinical judgment.

Interventions

The functional game used in the intervention was a mobile application designed to provide non-pharmacological treatment and cognitive training for individuals with MCI via smartphones or tablets. The goal of using this functional game was to enhance the cognitive function of participants in the experimental group compared to the control group by offering targeted cognitive training exercises. The application provides specific cognitive training activities (arithmetic training, attention training, reasoning, memory training, and observation training) in software format, presented through an interactive user interface (Table 1, Figs. 1 and 2). By providing repeated and progressively more challenging cognitive stimulation, the game helps activate brain functions and prevent or mitigate cognitive decline. Additionally, the application personalizes training based on user demographics such as age and sex, which enables the comparison of cognitive training performance with similar users. Real-time feedback, including program difficulty and ranking percentile, is provided to enhance user engagement and motivation for continuous cognitive training. To promote adherence to the prescribed frequency of at least three sessions per week, participants received systematic reminders via the application's built-in push notification system. Additionally, study staff conducted weekly phone calls to participants to check on game engagement, troubleshoot technical issues, and encourage continued participation. Game usage logs were automatically recorded and reviewed weekly to track individual adherence and ensure protocol compliance.

Table 1. Functional Game Content Overview.

Cognitive domain	Game	Description
Calculation	Calculation game 1	Enhances arithmetic and working memory by calculating and comparing which result is greater
	Calculation game 2	Tests speed and accuracy in performing basic arithmetic operations
	Calculation game 3	Challenges players to complete equations by selecting the correct numbers or symbols
	Calculation game 4	Requires remembering previous calculations while continuously adding or subtracting given numbers
	Calculation game 5	Trains real-life calculation skills by computing total prices, considering discounts and quantities
Attention	Attention game 1	Adapts to changing speeds and requires precise timing to tap accurately
	Attention game 2	Requires players to quickly assess obstacles and determine the correct moment to tap
	Attention game 3	Enhances concentration by making players quickly and accurately judge the direction a card should go
	Attention game 4	Players decide the priority of vehicles at an intersection to avoid accidents
	Attention game 5	Organizes parked cars in a way that ensures all vehicles can exit smoothly without obstacles
	Attention game 6	Improves focus by having players clap only when a specific target word appears
Reasoning	Reasoning game 1	Determines whether the image represents a left or right hand by visualizing spatial orientation
	Reasoning game 2	Enhances response inhibition by suppressing impulsive actions and following instructions to win based on known rules
	Reasoning game 3	Develops logical reasoning by deducing the correct path that satisfies the given game conditions
	Reasoning game 4	Trains the ability to quickly integrate directional cues in the mind to locate the correct position
Memory	Memory game 1	Trains memory by memorizing and inputting numbers displayed for a short duration, aiding real-life situations like entering verification codes
	Memory game 2	Enhances spatial memory by recalling specific locations in a given sequence
	Memory game 3	Enhances memory by recalling specific locations within a defined space
	Memory game 4	Enhances spatial memory by remembering the sequence of specific colors in a given space
	Memory game 5	Trains spatial memory by recalling specific shapes placed in designated areas
	Memory game 6	Enhances memory by recalling specific locations within a defined space
	Memory game 7	Improves sequential recall ability by remembering and recalling numbers in order
Observation	Observation game 1	Players identify a single shape with a slightly different color among multiple identical shapes as quickly as possible
	Observation game 2	Finds and inputs the missing number in a patterned sequence as quickly as possible
	Observation game 3	Quickly finds and selects all objects that match a given reference object among various displayed items
	Observation game 4	Identifies the single card that differs from the rest among multiple displayed cards as quickly as possible

Fig. 1. Functional game “Descartes” user interface 1.

(https://play.google.com/store/apps/details?id=com.braindr.brain&hl=ko-KR&pli=1).

Fig. 2. Functional game “Descartes” user interface 2.

(https://play.google.com/store/apps/details?id=com.braindr.brain&hl=ko-KR&pli=1).

Outcome measures

Attention domain of MMSE and MoCA

Human cognitive function operates under limited resources, meaning there is a restriction on processing all perceived sensory information. As a result, individuals selectively collect information related to their current tasks while filtering out unnecessary data. This selective processing is enabled by attention.¹⁶ Assessing attention in cognitive function testing is essential to comprehensively evaluating the overall cognitive state of an individual.¹⁷ Attention serves as a fundamental component of various cognitive processes and acts as an early indicator of cognitive decline. Research has identified diminished attention as a primary early warning sign of MCI and early-stage dementia.¹⁸,¹⁹,²⁰ Particularly, MMSE and MoCA are widely used tools for detecting such changes in their early stages.²¹ Lower attention scores can indicate difficulties in performing everyday tasks and suggest a potential decline in cognitive functioning. MMSE, commonly used for evaluating cognitive abilities in older adults, assesses attention using serial subtraction of 7 from 100 and spelling a word backward. In this study, the serial subtraction task was utilized. MoCA evaluates attention through multiple sub-tasks, including digit span (reciting numbers in reverse order), trail-making (connecting letters in sequence), and other tasks requiring sustained attention and recall. MoCA provides a broad assessment of attention by integrating various tasks that measure working memory and focus.

The total score for MMSE is 30, with attention contributing up to 5 points, while MoCA also has a total score of 30, with attention making up 6 points. The primary analysis in this study evaluated the differences in MMSE and MoCA attention scores between the experimental and control groups using a mixed-effects model. Additionally, we used a secondary analysis to compare the improvement rate in attention scores between groups. Attention improvement is defined as an increase of at least 1 point in the attention score from baseline to post-intervention. A two-sample z-test was conducted to compare the proportion of participants who showed attention improvement between the two groups. The inclusion of a proportion-based analysis is crucial, as mean-based evaluations might not reflect significant improvements in individual participants due to averaging effects. By conducting a proportion-based analysis, we were able to provide a more practical measure of cognitive enhancement.

MMSE assesses five domains: orientation, memory, attention, language, and visuospatial abilities, with attention comprising 16% of the total score. MoCA evaluates six domains: orientation, memory, attention, language, visuospatial abilities, and abstraction, with attention contributing 20% to the total score. Therefore, the improvement threshold for attention was conservatively set at 1 point. Supporting this approach, previous studies have suggested a 2-point improvement in total MMSE scores as clinically meaningful,²² while a 4-point increase in MoCA scores has also been proposed as a significant threshold.²³

CDR

CDR is widely utilized in clinical research and diagnosis tool to quantitatively evaluate dementia progression, detect early-stage dementia, and monitor changes in cognitive functioning over time.²⁴ CDR evaluates multiple components, including memory, perception, language, reasoning, and activities of daily living, and its primary advantage lies in its ability to measure the autonomy and functional independence of a patient, making it an essential tool for tracking cognitive decline. Studies have indicated that CDR is effective in assessing dementia progression²⁵ and that long-term fluctuations in CDR scores are correlated with changes in cognitive function.²⁶

Statistical analyses

All statistical analyses were performed using Stata software (version 18.5). The primary outcome variables were the MoCA and MMSE attention scores, while the secondary outcome variable was CDR.

For the primary outcome variables, a mixed-effects model was employed to analyze the effects of the intervention. The model included fixed effects for group (treatment vs. control), time (baseline and post-intervention), and their interaction (group×time) to evaluate whether changes in attention scores over time differed between the groups. Random effects were incorporated to account for variability at the individual level and for repeated measurements within subjects. Additionally, for the primary outcome variables, the improvement rates in attention scores between the treatment and control groups were compared using a two proportion z-test to assess differences in proportions of improvement. For the secondary outcome variable (CDR), a mixed-effects model and two-proportion z-test were also employed, using the same structure as for the primary outcomes, to examine group, time, and group×time interaction effects. Statistical significance was determined at a two-tailed p-value of <0.05.

RESULTS

Baseline characteristics

Of a total of 80 initial participants, two (one voluntary withdrawal, one loss of contact) were lost from the experimental group and seven (four voluntary withdrawal, three loss of contact) from the control group, resulting in data being collected and analyzed from participants (38 in the experimental group and 33 in the control group). The baseline characteristics of participants are summarized in Table 2. The experimental and control groups were compared in terms of demographic and clinical characteristics at baseline as described below:

Table 2. Baseline Characteristics of Study Participants (n=71).

Variable		Treatment group (n=38)	Control group (n=33)	p
Age (yr)		76.95±6.70	75.03±7.34	0.254
Sex, female		34 (89.47)	22 (66.66)	0.040
Education degree
	Under elementary school graduate	35 (92.10)	7 (21.2)
	Elementary school graduate	1 (2.63)	11 (33.33)
	Middle school graduate	0 (0)	3 (9.09)
	High school graduate	0 (0)	6 (18.18)
	University graduate	2 (5.26)	6 (18.18)	<0.001
MMSE		26.00±2.45	23.52±4.78	0.010
Attention of MMSE		3.39±1.65	2.97±1.70	0.291
	MMSE	26.00	23.52
	Orientation	9.42	8.94
	Memory	4.63	4.18
	Attention	3.39	2.97
	Language	7.74	6.73
	Spatial	0.82	0.70
MoCA		22.18±4.95	19.03±6.39	0.025
Attention of MoCA		4.05±1.61	4.06±1.75	0.984
	MoCA	22.18	19.03
	Spatial	3.53	2.64
	Language	5.03	4.45
	Memory	1.45	1.52
	Attention	4.05	4.06
	Abstract	1.42	1.00
	Orientation	5.82	5.03
CDR		0.49±0.14	0.61±0.39	0.084

MMSE, mini mental state examination; MoCA, Montreal cognitive assessment; CDR, clinical dementia rating.

Data are presented as mean±standard deviation or n (%).

Demographic characteristics

At baseline, the treatment group consisted of 38 participants, while the control group included 33 participants. The mean age of participants in the treatment group was 76.95±6.70 years, compared to 75.03±7.34 years in the control group, with no significant difference between the groups (p=0.254). Regarding sex, females were most common in both groups; however, a statistically significant difference was observed in the sex composition between the treatment and control groups (p=0.040). At baseline, the distribution of educational levels between the treatment and control groups was analyzed using a chi-square test for independence, which revealed significant differences in educational levels between the two groups (χ²=43.99, p<0.001). To account for differences in educational levels, 1 point was added to the MoCA scores of participants with fewer than 6 years of formal education, as per the standard MoCA scoring guidelines (Korea Central Dementia Center).

Baseline cognitive measures

Baseline cognitive measures, including attention scores of MMSE, MoCA, and CDR, were compared between the treatment and control groups. The mean attention score derived from the MMSE was 3.39±1.65 in the treatment group and 2.97±1.70 in the control group, showing no significant difference (p=0.291). Similarly, the mean attention score from the MoCA was 4.05±1.61 in the treatment group and 4.06±1.75 in the control group, with no significant difference observed (p=0.984). For the CDR, the mean score was slightly lower in the treatment group (0.49±0.14) compared to the control group (0.61±0.39); however, this difference was not statistically significant (p=0.084). The total MMSE score was 26.00 in the treatment group and 23.52 in the control group, showing a significant difference (p=0.010). Also, the total MoCA score was 22.18 in the treatment group and 19.03 in the control group, showing a significant difference (p=0.025) (Table 2).

Mixed-effects model results: attention of MMSE

A mixed-effects model was used to evaluate the effects of the intervention on attention scores measured by the MMSE. The model included fixed effects for group (treatment vs. control), time (baseline and post-intervention), and their interaction (group×time), along with random effects to account for variability at the individual level. The results are presented in Table 3. The main effect of time (baseline to post-intervention) was not statistically significant (p=0.319), indicating no significant overall change in MMSE attention scores across all participants. Similarly, the main effect of group (treatment vs. control) was also not significant (p=0.297), suggesting no significant difference in attention scores between the two groups at baseline. However, the interaction effect of group×time was also not significant (p=0.384), indicating that the improvement in MMSE attention scores over time did not differ significantly between the treatment and control groups (Fig. 3).

Table 3. Mixed-Effects Model Results for the Attention Domain of MMSE, MoCA, and CDR.

Effect		Coefficient	Std. error	Z-value	p
MMSE
	Intercept	2.970	0.298	9.965	<0.001
	Group (treatment)	0.425	0.407	1.043	0.297
	Time (post-intervention)	-0.273	0.274	-0.997	0.319
	Group×time Interaction	0.325	0.374	0.870	0.384
MoCA
	Intercept	4.061	0.286	14.222	<0.001
	Group (treatment)	-0.008	0.390	-0.020	0.984
	Time (post-intervention)	-0.303	0.260	-1.165	0.244
	Group×time Interaction	0.698	0.356	1.963	0.049
CDR
	Intercept	0.606	0.049	12.423	<0.001
	Group (treatment)	-0.119	0.067	-1.788	0.074
	Time (post-intervention)	0.045	0.020	2.255	0.024
	Group×time Interaction	-0.059	0.028	-2.127	0.033

MMSE, mini mental state examination; MoCA, Montreal cognitive assessment; CDR, clinical dementia rating.

Fig. 3. Mixed-effects model results for the attention domain of MMSE (A), MoCA (B), and CDR (C). Mean±SD by group and time. MMSE, mini mental state examination; MoCA, Montreal cognitive assessment; CDR, clinical dementia rating.

Mixed-effects model results: attention of MoCA

The mixed-effects model was also employed to analyze attention scores measured by the MoCA. The results, shown in Table 3, summarize the effects of group, time, and their interaction. The main effect of time was not statistically significant (p=0.244), suggesting no significant overall change in MoCA attention scores from baseline to post-intervention. Similarly, the main effect of group was also not found to be significant (p=0.984), indicating no significant difference in MoCA attention scores between the treatment and control groups at baseline. However, the interaction effect of group×time approached significance (p=0.050), suggesting a possible trend toward greater improvement in MoCA attention scores in the treatment group compared to the control group (Fig. 3).

Proportions of improvement: attention of MMSE

A two-proportion z-test was performed to evaluate the difference in the proportions of participants showing improvement in attention scores measured by the MMSE. Improvement was defined as an increase of at least 1 point in the attention score from baseline to post-intervention. The results are summarized in Tables 4 and 5.

Table 4. Mixed-Effects Model Results for MMSE, MoCA, and CDR.

Variables	Effect	Coefficient	Std. error	Z-value	p
MMSE (total)	Intercept	23.515	0.653	35.991	<0.001
MMSE (total)	group[T.1]	2.485	0.893	2.782	0.005
MMSE (total)	measure_time[T.1]	-0.333	0.450	-0.741	0.458
MMSE (total)	group[T.1]:measure_time[T.1]	0.702	0.614	1.142	0.253
MMSE (orientation to time/place)	Intercept	8.939	0.223	40.135	<0.001
MMSE(orientation to time/place)	group[T.1]	0.482	0.304	1.582	0.114
MMSE(orientation to time/place)	measure_time[T.1]	-0.152	0.218	-0.695	0.487
MMSE(orientation to time/place)	group[T.1]:measure_time[T.1]	0.152	0.298	0.508	0.611
MMSE (attention)	Intercept	2.970	0.298	9.965	<0.001
MMSE (attention)	group[T.1]	0.425	0.407	1.043	0.297
MMSE (attention)	measure_time[T.1]	-0.273	0.274	-0.997	0.319
MMSE (attention)	group[T.1]:measure_time[T.1]	0.325	0.374	0.870	0.384
MMSE (recall)	Intercept	4.182	0.200	20.940	<0.001
MMSE (recall)	group[T.1]	0.450	0.273	1.648	0.099
MMSE (recall)	measure_time[T.1]	-0.121	0.204	-0.595	0.552
MMSE (recall)	group[T.1]:measure_time[T.1]	0.490	0.278	1.760	0.078
MMSE (language)	Intercept	6.727	0.201	33.391	<0.001
MMSE (language)	group[T.1]	1.010	0.275	3.666	0.236
MMSE (language)	measure_time[T.1]	0.212	0.138	1.542	0.123
MMSE (language)	group[T.1]:measure_time[T.1]	-0.107	0.188	-0.568	0.570
MMSE (visual construction)	Intercept	0.697	0.079	8.857	<0.001
MMSE (visual construction)	group[T.1]	0.119	0.108	1.105	0.269
MMSE (visual construction)	measure_time[T.1]	0.000	0.081	0.000	>0.999
MMSE (visual construction)	group[T.1]:measure_time[T.1]	-0.158	0.111	-1.421	0.155
MoCA (total)	Intercept	19.030	0.956	19.917	<0.001
MoCA (total)	group[T.1]	3.154	1.306	2.415	0.016
MoCA (total)	measure_time[T.1]	-0.394	0.616	-0.639	0.523
MoCA (total)	group[T.1]:measure_time[T.1]	-1.264	0.842	-1.501	0.133
MoCA (visuospatial/executive)	Intercept	2.636	0.266	9.922	<0.001
MoCA (visuospatial/executive)	group[T.1]	0.890	0.363	2.450	0.014
MoCA (visuospatial/executive)	measure_time[T.1]	0.091	0.228	0.399	0.690
MoCA (visuospatial/executive)	group[T.1]:measure_time[T.1]	-1.670	0.311	-5.367	<0.001
MoCA (attention)	Intercept	4.061	0.286	14.222	<0.001
MoCA (attention)	group[T.1]	-0.008	0.390	-0.020	0.984
MoCA (attention)	measure_time[T.1]	-0.303	0.260	-1.165	0.244
MoCA (attention)	group[T.1]:measure_time[T.1]	0.698	0.356	1.963	0.049
MoCA (language)	Intercept	4.455	0.239	18.632	<0.001
MoCA (language)	group[T.1]	0.572	0.327	1.750	0.080
MoCA (language)	measure_time[T.1]	0.000	0.187	0.000	>0.999
MoCA (language)	group[T.1]:measure_time[T.1]	-0.053	0.256	-0.206	0.837
MoCA (abstraction)	Intercept	1.000	0.136	7.331	<0.001
MoCA (abstraction)	group[T.1]	0.421	0.186	2.258	0.024
MoCA (abstraction)	measure_time[T.1]	0.000	0.142	0.000	>0.999
MoCA (abstraction)	group[T.1]:measure_time[T.1]	0.053	0.194	0.271	0.786
MoCA (delayed recall)	Intercept	1.515	0.258	5.876	<0.001
MoCA (delayed recall)	group[T.1]	-0.068	0.352	-0.192	0.847
MoCA (delayed recall)	measure_time[T.1]	-0.273	0.262	-1.041	0.298
MoCA (delayed recall)	group[T.1]:measure_time[T.1]	-0.017	0.358	-0.047	0.963
MoCA (orientation)	Intercept	5.030	0.180	27.940	<0.001
MoCA (orientation)	group[T.1]	0.785	0.246	3.192	0.001
MoCA (orientation)	measure_time[T.1]	0.061	0.151	0.400	0.689
MoCA (orientation)	group[T.1]:measure_time[T.1]	-0.192	0.207	-0.928	0.353
CDR	Intercept	0.606	0.049	12.423	0.000
CDR	group[T.1]	-0.119	0.067	-1.788	0.074
CDR	measure_time[T.1]	0.045	0.020	2.255	0.024
CDR	group[T.1]:measure_time[T.1]	-0.059	0.028	-2.127	0.033

MMSE, mini mental state examination; MoCA, Montreal cognitive assessment; CDR, clinical dementia rating.

Table 5. Two-Proportion Z-Test Analysis of Attention Scores from MMSE, MoCA, and CDR.

Group		Improved count	Total count	Proportion (%)	Z-statistic	p
MMSE					0.050	0.615
	Control	7	33	18.18
	Treatment	10	38	23.68
MoCA					-2.970	0.003
	Control	6	33	18.18
	Treatment	16	38	42.11
CDR					0.939	0.174
	Control	0	33	0.00
	Treatment	1	38	2.63

MMSE, mini mental state examination; MoCA, Montreal cognitive assessment; CDR, clinical dementia rating.

The analysis did not reveal a significant difference in the proportions of improvement between the treatment and control groups (p=0.615). These results suggest that the intervention in the treatment group was not associated with a higher likelihood of improvement in MMSE attention scores compared to the control group.

Proportions of improvement: attention of MoCA

A two-proportion z-test was performed to evaluate the difference in the proportions of participants showing improvement in attention scores measured by the MoCA. Improvement was defined as an increase of at least 1 point in the attention score from baseline to post-intervention. The results are summarized in Tables 4 and 5.

The analysis revealed a significant difference in the proportions of improvement between the treatment and control groups. In the treatment group, 42.11% of participants showed improvement in MoCA attention scores, compared to 18.18% in the control group. The two-proportion z-test yielded a z-statistic of -2.97 and a p-value of 0.003, indicating that the difference was statistically significant (p<0.05).

These results suggest that the intervention in the treatment group was associated with a higher likelihood of improvement in MoCA attention scores compared to the control group. This finding supports the efficacy of the intervention in enhancing cognitive attention as measured by the MoCA.

Mixed-effects model results: CDR

A mixed-effects model was employed to evaluate the effects of the intervention on the CDR. The model included fixed effects for group (treatment vs. control), time (baseline and post-intervention), and their interaction (group×time), along with random effects to account for variability at the individual level. The results are presented in Table 3 and Fig. 3.

The main effect of time (baseline to post-intervention) was statistically significant (p=0.024), indicating that CDR scores significantly changed over time across all participants. However, the main effect of group (treatment vs. control) was not significant (p=0.074), suggesting no significant difference in baseline CDR scores between the two groups. The group×time interaction effect was, however, statistically significant (p=0.033), indicating that the change in CDR scores over time differed significantly between the treatment and control groups. This suggests a differential impact of the intervention on CDR scores.

The analysis revealed a significant time effect (p=0.024), indicating that CDR scores increased over time across all participants, suggesting a general trend of cognitive decline. However, the significant group×time interaction effect (p=0.033) indicates that the rate of increase in CDR scores over time differed between the treatment and control groups. Specifically, the treatment group demonstrated a smaller increase in CDR scores than the control group, implying that the intervention may have slowed cognitive decline.

These findings suggest that the intervention had a protective effect on cognitive function as measured by the CDR by reducing the extent of decline in the treatment group compared to the control group. However, the main group effect was not significant (p=0.074), indicating that baseline CDR scores were not significantly different between the two groups.

Proportions of improvement: CDR

A two-proportion z-test was performed to evaluate the difference in the proportions of participants showing improvement in attention scores measured by the CDR. Improvement was defined as a decrease of at least 0.5 points in the CDR score from baseline to post-intervention. The results are summarized in Tables 4 and 5. The analysis did not reveal a significant difference in the proportions of improvement between the treatment and control groups (p=0.174).

DISCUSSION

The goal of this study was to evaluate the cognitive enhancement effects of a mobile-based, game-driven intervention on the attention of older adults with MCI. Our findings provide insights into the efficacy of the use of functional games in improving specific cognitive domains, particularly attention, as measured by MoCA, and overall cognitive function, as assessed through CDR. The detailed implications of these findings are as follows.

A comparison of MMSE attention scores before and after the intervention showed no statistically significant difference between the experimental and control groups. Additionally, the two-proportion z-test (defining improvement as an increase of ≥1 point) indicated a higher improvement rate in the experimental group, although the difference was not statistically significant. These findings suggest that the use of functional games may have a limited impact on attention as measured by MMSE (Table 4).

Several factors may explain these findings. First, while MMSE is widely used for general cognitive screening, it may not be sensitive enough to detect subtle changes in specific cognitive domains such as attention. Second, ceiling effects might have contributed to false negatives, particularly for participants who had relatively high baseline scores.²⁷ Therefore, supplementary assessment tools may be needed to accurately measure attention improvements. These findings are consistent with previous literature suggesting that the MMSE lacks sensitivity in detecting subtle changes in specific cognitive domains such as attention. Nasreddine, et al.²³ developed the MoCA specifically to address these limitations, particularly in the assessment of attention and executive function. Similarly, Arevalo-Rodriguez, et al.²⁸ noted that the MMSE has limited performance in detecting early cognitive decline in individuals with MCI, especially in attention-related domains.

In contrast to MMSE, MoCA attention scores showed significant improvement in the experimental group compared to the control group. A two-proportion z-test confirmed a statistically significant improvement rate (≥1 point increase) in the experimental group. These findings suggest that the functional games used in this study have the potential to enhance attention in elderly individuals living with MCI (Table 4).

The differing results between MMSE and MoCA highlight the importance of sensitivity and specificity in cognitive assessment tools. MoCA is designed to evaluate executive function and attention more comprehensively, making it more effective at detecting subtle improvements that MMSE may overlook. Previous research has reported that MoCA is better suited than MMSE for detecting early cognitive decline, particularly in MCI.²⁹,³⁰ This suggests that, when designing cognitive interventions targeting specific cognitive functions such as attention, using more sensitive assessment tools like the MoCA is critically important. While MMSE remains a practical screening tool due to its simplicity, its limitations may lead to underestimation of intervention effects. In contrast, the comprehensive approach characteristic of the MoCA in evaluating cognitive domains, including attention and executive function, makes it a more sensitive tool for assessing targeted cognitive interventions.

CDR scores showed a significant decrease in the experimental group compared to the control group, indicating overall cognitive improvement in participants exposed to a functional game intervention (Table 3). These findings suggest that the use of functional games may have broad positive effects on cognitive functioning beyond attention alone. The interactive nature of functional games likely contributed to improved neuroplasticity by engaging multiple cognitive processes including memory, problem-solving, and spatial abilities.³¹,³² This finding aligns with previous research providing evidence that multidomain cognitive interventions can lead to meaningful improvements in overall cognitive status.³³,³⁴ Thus, functional games may be a valuable tool for supporting cognitive function in elderly individuals with MCI.

The findings of this study provide practical implications for cognitive rehabilitation in older adults living with MCI. The use of functional games was found to be effective in improving attention, as measured by the MoCA, and overall cognitive function, as reflected in CDR scores. These results suggest that interactive, engaging, and stimulating activities such as functional games should be integrated into cognitive training programs for older adults living with MCI.

However, the lack of significant improvement in MMSE scores highlights the importance of selecting appropriate assessment tools when evaluating intervention efficacy. Future studies should carefully consider the most suitable cognitive assessment tools based on the targeted cognitive domains. Additionally, larger sample sizes and longitudinal studies are necessary to further confirm the long-term effects of functional game use on cognitive function.

This study has several inherent limitations. First, there was a significant difference in total MMSE and MoCA scores at baseline between the experimental and control groups, with the participants in the experimental group scoring higher. However, as there was no difference between the two groups in the attention domain despite the overall higher MMSE and MoCA scores in the experimental group, this finding suggests that the attention function of the experimental group participants was relatively lower compared to other cognitive functions. The high percentage of individuals with low educational levels may indicate that attention function, including calculating ability, may have been more impaired compared to other cognitive domains. This highlights the need to apply functional games to improve attention, as the study confirmed the potential for attention enhancement in the MoCA attention test when functional games were used by patients with relatively impaired attention. Second, the relatively small sample size may have limited the statistical power to detect significant differences, particularly in MMSE results. Third, the intervention duration may have been too short to capture long-term cognitive improvements. Fourth, the study population consisted exclusively of older adults living with MCI. Fifth, CDR is a tool to measure the overall stage of dementia or cognitive impairment. Therefore, caution is warranted when interpreting the finding that the CDR of the experimental group was less worsened than that of the control group after 3 months of applying the functional game to the study participants. Upon reviewing the subdomains, some improvement was observed in judgment and problem-solving abilities, while no differences were found in memory, orientation, and activity. Given that the participants had MCI with slow progression, it is presumed that improvement occurred in the areas related to stimulation of the functional game after applying the intervention. Sixth, the absence of a placebo application in the control group may present a limitation in the research design. This study compared the experimental group (drug+functional game) with the control group (drug only) due to practical constraints, including a short research period that made it difficult to develop a placebo application. However, a placebo application will be used in the control group in future confirmatory clinical trials. Finally, although this study was conducted as a RCT, the research protocol was not registered in a public clinical trial registry such as ClinicalTrials.gov prior to participant enrollment. Nonetheless, the full study protocol and SAP were established before the trial began and were approved by the institutional review board. We acknowledge the importance of prospective registration for enhancing research transparency and will ensure protocol registration in future trials.

In conclusion, this study provides evidence of the effectiveness of a mobile-based game-driven program that was customized to meet the needs of older adults living with MCI. These functional games that offer individualized cognitive programs are instrumental in improving the cognitive function of attention among older adults with MCI. Designing and implementing a variety of mobile-based cognitive training programs has the potential to significantly contribute to supporting and improving the cognitive functioning of older adults living with MCI.

AVAILABILITY OF DATA AND MATERIAL

The data and materials supporting the findings of this study are available upon reasonable request.