Inclusion and exclusion criteria

Literate older adults capable of understanding and following instructions were included. Individuals were excluded if they were illiterate, displaced, or institutionalized, as well as those with uncorrected auditory and/or visual impairments. Also excluded were participants with vertigo, nausea, dizziness, balance issues, or disabling motor difficulties requiring assistive devices (e.g., crutches, canes, or wheelchairs). Additionally, individuals with moderate or severe cognitive impairment-verified through the administration of the Mini-Mental State Examination (MMSE; ^{Escobar et al., 1999})-were excluded.

Procedure

This study was approved by the Ethics Committee of Universidad El Bosque under code PCI 2023-0026. All ethical requirements for research involving human subjects established by Colombian legislation were met, in accordance with the guidelines of the Declaration of Helsinki and Resolution 8430 of 1993 on human subjects research (^{Ministerio de Salud, 1993}), with a minimal risk classification.

Participants over 50 years of age enrolled in the active aging program of the Compensar family compensation fund in Bogotá, Colombia, were invited to participate via word of mouth. Those who voluntarily agreed to participate, after signing informed consent, were scheduled for individual appointments at Compensar facilities, where they were assessed by final-semester psychology students under the supervision of licensed psychologists who were previously trained in the administration and interpretation of the evaluation protocol.

Each participant was assigned a numeric code to protect their identity. Group randomization was conducted using the free online platform Sealed Envelope. The intervention then began, with its structure detailed in Table 1. Based on the intervention protocols described by ^{Lund et al. (2020a}), which recommend between 8 and 12 sessions, this study implemented 10 MotoTiles training sessions for the treatment group. Each session was led by at least one psychologist from the research team, conducted once per week, and lasted 60 minutes.

The following outlines the phase-specific structure of the intervention to ensure methodological replicability:

A detailed description of the games, the cognitive functions stimulated, and the motor domains are presented in Table 1, and the overall procedure is schematically illustrated in Figure 1.

Control Group Intervention Protocol: The control group was subjected to a low-intensity cognitive training protocol, which was designed to control the effects of attention and social interaction. The protocol consisted of a total of four standardized sessions, administered once per week. The intervention was conducted using the computerized platform CogniFit, a commercially available program focused on training specific cognitive functions (such as attention, working memory, and processing speed). During each session, participants individually performed the exercises proposed by the platform, remaining seated in front of a computer for a period of 45 minutes.

Figure 1: Flow diagram of the study procedure phases 

Table 1: Games Used in the Cognitive Training Sessions with MotoTiles 

^Note: Information taken from the Moto Tiles Manual (^{Entertainment Robotics, 2014}); for videos and more details, visit <https://www.Moto-tiles.com/>.

On the other hand, the waitlist control group was offered a simple and low-intensity training program (four sessions) based on the use of the computerized program CogniFit; participation was voluntary. Similarly, after the post-intervention assessment was completed for both groups, participants in the control group were offered additional training sessions using the Moto protocol, with voluntary participation.

Instruments

Demographic and Inclusion Criteria Questionnaire. This questionnaire was designed by the research team for the purpose of collecting general information such as gender, age, and educational level. It also included questions related to potential cognitive and/or motor conditions that could hinder participation in the study.

Cognitive Change Complaint Questionnaire (QCC). The QCC, developed and validated by ^{Salazar and Maldonado (2020}), assesses the subjective perception of cognitive changes across three dimensions: (1) cognitive, (2) global functioning, and (3) emotional. It consists of 26 items and has a cutoff score of 51. Its reliability was determined using Cronbach’s alpha coefficient, yielding high internal consistency (α = .973), along with 96.8 % sensitivity and 96.1 % specificity.

^{Rey Complex Figure (copy and recall). This test was developed by André Rey in 1941} and later standardized by ^{Osterrieth in 1944}. It evaluates visual perception and memory using 18 elements scored for correct location and accuracy (Osterrieth, 1944). The test was validated in Colombia by ^{Arango-Lasprilla et al. (2022}), showing a reliability coefficient of 0.85 for the copy phase and .97 for the recall phase.

Montreal Cognitive Assessment (MoCA). The MoCA is a brief screening tool that is both sensitive and specific for detecting mild cognitive impairment and dementia in older adults. Developed by ^{Nasreddine et al. in 2005}, it assesses memory, attention, language, visuospatial ability, task execution, and abstract thinking. It consists of 30 items. Although the general cutoff score is < 26, for this study a threshold of < 24 was adopted based on the Colombian validation by ^{Pedraza et al. (2016}), which reported moderate test-retest reliability (.62) and good internal consistency (.85).

Ineco Frontal Screening (IFS). The IFS is a neuropsychological battery developed by the Institute of Cognitive Neurology (^{Torralva et al., 2009}) to assess executive functions. It comprises eight subtests with a maximum score of 30 points. The Colombian validation by ^{Zapata-Zabala et al. (2019}) was used in this study, which establishes a cutoff of < 23 to identify frontal dysfunction. The test demonstrated strong discriminatory power under the ROC curve (AUC = 0.93, p < .001), with 80 % sensitivity and 91 % specificity.

Symbol Digit Modalities Test (SDMT). Developed by ^{Smith in 1973} and validated in Colombia by ^{Arango-Lasprilla et al. (2015}), the SDMT assesses sustained attention, visual scanning, and perceptual speed. The test presents nine geometric figures, each paired with a number, and participants are required to correctly match them as quickly as possible. Although originally administered in 90 seconds, the time was extended to 120 seconds in this study, following the Colombian validation by Arango-Lasprilla et al. (2022). Among older adults, reliability ranges from .263 to .351. The test shows good convergent validity with the WISC-R and Stroop tests, and a test-retest correlation of approximately .80 in healthy populations.

Trail Making Test (TMT). Created by ^{Reitan and Wolfson (1985}) and validated in Colombia by ^{Arango-Lasprilla et al. (2015}), the TMT assesses visual attention, processing speed, cognitive flexibility, and executive functions. It is divided into two parts: TMT-A (sustained attention) and TMT-B (alternating attention with executive demands). The test demonstrates high reliability: r = .99 for TMT-A and r = .93 for TMT-B. Construct validity was confirmed via factor analysis in the Colombian validation by Arango-Lasprilla et al. (2022).

Instrumented Timed Up and Go (iTUG). Functional mobility and balance were assessed using the iTUG test, which employed inertial measurement units (IMUs) from the Xsens brand. These sensors, placed on eight body regions, recorded acceleration, rotation, and magnetic field data. For kinematic analysis, only the data from the sensor positioned on the pelvis were used due to its proximity to the center of mass and the quality of recorded signals.

Statistical analysis

Cognitive data were analyzed using ^{Jamovi software (version 2023)}. Prior to inferential analysis, the normality of distributions was tested using the Shapiro-Wilk test, appropriate for samples with over 50 participants. Results indicated that cognitive variables followed a non-parametric distribution.

Descriptive statistics (mean, standard deviation, frequencies, and percentages, as appropriate) were used to describe the sample's demographic characteristics. To explore relationships between key variables and control for potential sources of variability, linear mixed models were applied, using sociodemographic variables as covariates. The models included mean scores for each test, along with significance levels associated with treatment effects, time of measurement, and interaction. This approach enabled the incorporation of both fixed effects (related to the study’s main independent variables) and random effects (associated with inter- and intra-individual variability), providing a robust and flexible statistical framework for the analysis of longitudinal and/or clustered data.

The use of linear mixed models allowed for the precise evaluation of intra- and inter-group differences, effectively controlling sample heterogeneity and optimizing the interpretation of results in line with the study objectives. All analyses were conducted using a statistical significance level of p < .05.

Motor data were processed using specialized motion analysis software. Variables such as total test time and gait characteristics (e.g., velocity, cadence, number of steps) were assessed, with outbound and return phases analyzed separately in order to build a detailed profile of gait patterns. Linear mixed models (LMMs) were also used to compare pre- and post-intervention measurements, accounting for repeated measures and individual variability (^{UCLA, 2025}). To control Type I error in multiple comparisons, the Holm-Bonferroni correction was applied. Compared to the traditional Bonferroni method, this approach is less conservative and better preserves statistical power in small samples (^{Editverse, 2025}).

Cognitive results

Table 2 presents the average scores obtained on the cognitive tests before and after the intervention for both groups. A slight improvement was observed in the post-intervention scores of the treatment group, particularly in the MoCA, INECO, and the memory component of the Rey Complex Figure test. In addition, a reduction in subjective cognitive complaints was noted, as measured by the Cognitive Change Complaint Questionnaire (QCC).

Table 2: Average Cognitive Scores in Pre- and Post-Intervention Assessments for Both Groups 

^Note: CG: waitlist control group; TG=treatment group; QCC: Cognitive Change Complaint Questionnaire; TMTA and TMTB: Trail Making Test A and B; ReyCopy: Rey Complex Figure (copy); ReyRecall: Rey Complex Figure (recall); MoCA: Montreal Cognitive Assessment Test.

The significance analysis (Table 3) showed that the intervention group obtained an estimated change score of -0.9166 on the MoCA test, suggesting a slight improvement in the post-test compared to the pre-test. However, the 95 % confidence interval (-3.5651 to 1.7320) included the null value, indicating that the difference did not reach statistical significance. Among the covariates analyzed, educational level showed a significant association with cognitive performance, suggesting that a higher level of education is related to better results on cognitive tests.

In the Trail Making Test A (TMT-A), statistical significance was not reached (p = .095), although a trend was observed that may be relevant in studies with a larger sample size. For the Trail Making Test B (TMT-B), education again showed a significant association (p = .035). In the Symbol Digit Modalities Test (SDMT), both age (p < .001) and education (p < .001) were significantly associated with better performance, highlighting the influence of these factors on alternating attention and processing speed.

Table 3: Correlation Between Study Variables and the MoCA, TMT A and B, and SDMT Tests 

^Note: MoCA: Montreal Cognitive Assessment Test; TMA A and B: Trail Making Test A and B; SDMT: Symbol Digit Modalities Test. *p <= .05

Regarding the Rey Complex Figure (Table 4), the results indicated that both age (p = .017) and education (p < .001) were significantly associated with performance in the delayed recall phase. In the copy phase, only education showed a significant relationship (p = .012). These findings reinforce the importance of educational level and age in visuospatial and memory task performance.

Table 4: Correlation Between Study Variables and the Rey Complex Figure Test (Copy and Delayed Recall) 

*p <= .05

Taken together, the cognitive results suggest that the intervention led to slight improvements in the experimental group, although most tests did not reach statistical significance. Education emerged as a key factor associated with better cognitive performance, while age primarily influenced tasks involving memory and processing speed. These observations are consistent with previous literature highlighting the moderating role of sociodemographic variables in cognitive performance.

Motor results

Table 5 summarizes the results of the pre- and post-intervention motor assessments for both groups, analyzed in a subsample of 28 participants (7 in the control group and 21 in the experimental group). In the experimental group, statistically significant differences were identified in two key indicators: gait speed (p = .030) and total walking time (p = .027).

These changes suggest a modification in participants' motor strategy following the intervention, possibly aimed at optimizing stability or adapting to neuromuscular demands. Although variables such as step length, number of strides, and stride length did not reach statistical significance, a trend toward a more conservative gait pattern was observed in the experimental group.

In contrast, the control group did not show significant differences between pre- and post-intervention measurements, which supports the hypothesis that the changes observed in the experimental group are attributable to the intervention.

Table 5: Pre- and Post-Intervention Comparison of Motor Outcome Variables 

^Note: TG: treatment group; CG: waitlist control group.

The motor results indicate that the intervention had a positive and significant impact on gait speed and walking time in the experimental group. These changes may reflect a favorable adaptation in motor strategy aimed at maintaining stability and safety during gait key aspects in populations vulnerable to motor impairments.

Cognitive performance

In the cognitive domain, a slight improvement was observed in MoCA scores, as well as in the INECO test and the memory item of the Rey Complex Figure, alongside a reduction in subjective cognitive complaints. Although the difference in MoCA was not statistically significant, other tests such as TMT-A showed only a trend toward statistical significance (p = .095), which could be attributed to the limited sample size and insufficient statistical power to detect subtle changes. In terms of cognitive functions, the improvements observed in these tests are attributed to a positive impact on general cognition, sustained attention, visuoconstructional processes, and reduced reaction time.

These findings are consistent with previous studies reporting modest positive effects of cognitive interventions in older adults, particularly when training frequency and duration are moderate (^{Liu et al., 2018}). These results also align with previous MotoTiles studies reporting selective improvements in attention, visuospatial processing, and reaction speed (Liu et al., 2018), particularly when interventions are implemented with similar session frequency and duration. Furthermore, education and age were found to significantly influence cognitive performance, reinforcing the concept of cognitive reserve and the importance of sociodemographic factors in the response to cognitive interventions (^{Johnson et al., 2019}; ^{Smith et al., 2020}). This pattern aligns with literature on age-related cognitive trajectories. ^{Salthouse (2019}) highlights that processing speed, attentional capacity, and memory typically decline with advancing age. In this context, even modest gains, as observed in the present study, may be clinically meaningful, as they counteract the expected downward trend associated with normal aging. Moreover, in aging populations, the mere maintenance of baseline cognitive performance is considered a positive outcome, as it indicates resilience against the progressive decline typically observed over time. This perspective reinforces the relevance of the present findings, even when improvements are modest or non-significant from a statistical standpoint.

Similarly, ^{Arokiaraj et al. (2020}) reported that computerized cognitive training in healthy older adults tends to produce small-to-moderate improvements, particularly in processing speed and attentional domains, when training intensity and duration are comparable to those used in the present study. Likewise, meta-analytic findings by ^{Chan et al. (2024}) indicate that technology-based cognitive training yields modest yet meaningful effects on memory and other cognitive functions in older adults with cognitive impairment, further supporting the pattern of selective gains observed here. Comparable patterns have also been observed in studies where gains in memory or executive functioning did not reach statistical significance, often due to the cognitive complexity of the tasks involved or the limited familiarity participants had with game-based activities (^{Liu et al., 2018}).

Motor performance

Regarding motor outcomes, statistically significant differences were identified in gait speed (p= .030) and total walking time (p = .027) in the experimental group, while other variables such as step length and number of strides showed non-significant trends. These findings suggest that participants modified their motor strategies, possibly adopting a more conservative gait pattern as a compensatory mechanism to maintain stability or respond to new neuromuscular demands following the intervention. The literature supports that both physical and cognitive interventions can induce functional adaptations in gait patterns, especially in older adults with biomechanical or neuromotor alterations (^{Eggenberger et al., 2015}). In particular, Eggenberger et al. (2015) demonstrated that simultaneous cognitive-motor training can enhance gait velocity and stability, a result that closely mirrors the improvements seen in the present study and further reinforces the value of dual-task-oriented approaches like MotoTiles. Thus, previous MotoTiles interventions have shown improvements in balance, coordination, and fall-prevention-related motor parameters (^{Ehrari et al., 2020}; ^{Lund et al., 2020a}), which supports the functional relevance of the gait changes observed here. The absence of significant changes in the control group further supports the hypothesis that the observed modifications are directly related to the intervention.

Finally, the clinical and social relevance of this study and its findings should be highlighted. Cognitive and motor training using MotoTiles represents an innovative and accessible strategy to promote overall health in older adults, contributing to both the prevention and slowing of cognitive and motor decline associated with aging. Clinically, the findings suggest that this type of intervention may help preserve executive functions, attentional processes, and processing speed (as reflected in reaction times), as well as improve key motor parameters such as gait speed factors that translate into greater autonomy and reduced risk of dependency and falls. Socially, the implementation of interactive and group-based programs like MotoTiles fosters active participation, a sense of belonging, and social integration crucial elements for combating isolation and improving quality of life in old age. Moreover, the adaptability of the intervention to different educational levels and functional capacities broadens its applicability and potential public health impact, helping to reduce the socioeconomic burden associated with disability and functional decline in aging populations.

Limitations

The current study identified several methodological limitations that must be considered when interpreting the results and that define future research lines. The main constraint lies in the sample characteristics, which was reduced, predominantly female, and defined by strict exclusion criteria. This sample homogeneity compromises the external validity of the findings and their generalization to more heterogeneous populations, including individuals with lower educational levels (illiteracy), moderate/severe cognitive impairment, or motor and balance difficulties.

From an analytical perspective, although linear mixed models were used to control for covariates, the significant influence observed from sociodemographic variables such as age and educational level on cognitive performance constitutes a factor that moderates the interpretation of the pure intervention effect.

Regarding the training protocol, the low intensity and duration may have limited the statistical power to detect significant changes. Furthermore, the absence of long-term follow-up prevents the determination of the temporal persistence of the observed positive effects. Finally, the use of conventional assessment instruments, based on global scores, may not have captured all the subtle or specific changes induced by the multimodal MotoTiles intervention.

Future Directions

To optimize the impact of future research and strengthen empirical evidence, the following are recommended: Increase the sample size and ensure greater sociodemographic diversity to enhance external validity and the generalizability of results. Increase the frequency and/or duration of training sessions to determine the optimal dose required to generate significant effects. Incorporate a more exhaustive analysis of cognitive and motor reserve at baseline. This will allow for the use of more robust statistical models or designs that facilitate better group stratification and matching, more accurately isolating the causal effect of the MotoTiles intervention. Implement long-term follow-up to evaluate the stability and persistence of the training effects. Utilize more sensitive and specific measurement instruments that go beyond global scores to capture subtle changes in executive functions and kinematic parameters.