JMIR Publications

ABSTRACT

Background:

The ankle joint is central to postural adjustment and daily movement; age-related decline in ankle function directly affects balance and mobility. Accessible, repeatable ankle control methods remain scarce for older adults. Serious games offer a potential alternative for exercise and high-resolution movement monitoring. However, many systems emphasize overall task success or conventional objective outcomes, while detailed analysis of movement signals during dynamic gameplay remains limited.

Objective:

This study aimed to develop Balance Wood, a serious game for interactive digital analysis of ankle range of motion. We examined the system’s feasibility and measurement sensitivity in capturing behavioral adaptations in older and younger adults, establishing an interpretable, signal-based analysis framework.

Methods:

Balance Wood is a bilateral ankle-control system where two tilting foot devices generate a combined control angle for a bucket-control serious game. We included 40 participants (20 older and 20 younger adults) who completed three consecutive trials of varying difficulty. From the recorded tilt-angle signals, five main trial-level outcomes were derived for inferential analysis: typical forward amplitude, typical backward amplitude, forward holding proportion, backward holding proportion, and overspeed proportion. Effective proportions of forward and backward flexion-extension were calculated to measure movement patterns descriptively. Repeated-measures analyses of variance were performed separately for both age groups. Sphericity was assessed using the Mauchly test, applying Greenhouse-Geisser correction when required. To address outcome-level multiplicity for the five outcomes per group, a Bonferroni-corrected threshold of P<.01 was used, with effect sizes interpreted alongside uncorrected P values.

Results:

In older adults, repeated-measures analysis of variance revealed nominally significant trial effects (uncorrected P<.05) for typical forward amplitude (F_2,38=4.169, P=.023, η_p^2=.180) and typical backward amplitude (F_1.373,26.084=4.338, P=.036, η_p^2=.186; Greenhouse-Geisser corrected). Despite large effect sizes, neither survived the Bonferroni-corrected threshold of P<.01. Forward holding proportion (F_2,38=3.122, P=.056, η_p^2=.141) and backward holding proportion (F_2,38=2.563, P=.090, η_p^2=.119) did not reach nominal statistical significance. Overspeed proportion remained stable (F_2,38=0.416, P=.663, η_p^2=.021). Post hoc testing indicated the most evident pairwise change was an increase in typical backward amplitude from Trial 1 to Trial 2, demonstrating system sensitivity to task-induced adaptations. In younger adults, no inferential outcomes showed significant trial effects, suggesting a ceiling effect or greater movement reserve. Descriptively, older adults showed increasing typical amplitudes and a directionally asymmetric profile of effective flexion-extension, whereas younger adults exhibited numerically larger amplitudes and a more directionally balanced profile.

Conclusions:

In older adults, repeated-measures analysis of variance revealed nominally significant trial effects (uncorrected P<.05) for typical forward amplitude (F_2,38=4.169, P=.023, η_p^2=.180) and typical backward amplitude (F_1.373,26.084=4.338, P=.036, η_p^2=.186; Greenhouse-Geisser corrected). Despite large effect sizes, neither survived the Bonferroni-corrected threshold of P<.01. Forward holding proportion (F_2,38=3.122, P=.056, η_p^2=.141) and backward holding proportion (F_2,38=2.563, P=.090, η_p^2=.119) did not reach nominal statistical significance. Overspeed proportion remained stable (F_2,38=0.416, P=.663, η_p^2=.021). Post hoc testing indicated the most evident pairwise change was an increase in typical backward amplitude from Trial 1 to Trial 2, demonstrating system sensitivity to task-induced adaptations. In younger adults, no inferential outcomes showed significant trial effects, suggesting a ceiling effect or greater movement reserve. Descriptively, older adults showed increasing typical amplitudes and a directionally asymmetric profile of effective flexion-extension, whereas younger adults exhibited numerically larger amplitudes and a more directionally balanced profile. Clinical Trial: Findings demonstrate the feasibility and sensitivity of Balance Wood for digital analysis of ankle control. The system successfully captured behavioral adaptations and distinct movement strategies responding to changing task demands. The signal-based framework shows significant potential for continuous monitoring and detailed profiling of directional amplitude, near-limit holding behavior, and speed regulation during gameplay.