Analysis of Spatiotemporal Features in a Virtual Navigation Game Across Different Age Groups: Quantitative Research

With rapid urbanization, the proliferation of densely arranged buildings and increasingly homogeneous architectural designs has made disorientation and navigation difficulties more common, especially for older adults. Meanwhile, advances in virtual reality technology now allow researchers to create highly immersive navigation games, offering opportunities for assessing cognitive abilities and examining how environmental factors shape navigation behavior.

This study aimed to design a virtual reality-based navigation game capable of assessing cognitive abilities through navigation behavior and quantitatively examining how environmental configurations influence navigation patterns in different age groups.

We designed a virtual goal-directed navigation game and recruited 2 groups, younger adults (n=18) and older adults (n=21), to complete identical wayfinding tasks. Before the formal experiment, participants completed cognitive assessments and received training. To characterize navigational behavior, k-means clustering was applied to classify navigation states and extract behaviorally meaningful navigation measurements, which were then examined for correlations with cognitive test scores. To quantify the effects of environmental structure, space syntax analysis was conducted to calculate line-based and grid-based experienced metrics for each participant, and their associations with navigation performance were examined. Additionally, between-group differences in navigation performance and experienced metrics were evaluated across age groups.

Our results revealed that navigation behavior performance, particularly navigation efficiency, was significantly influenced by cognitive abilities and was strongly associated with several cognitive tests: the Montreal Cognitive Assessment (r=0.495, P=.04), Trail Making Test Part A (r=-0.761, P=.001), and the Mental Rotation Test (r=0.848, P<.001). In terms of environmental influences, experienced axial integration (EAI) and experienced visual integration (EVI) demonstrated significant age-related differences: EAI (z=-2.43, P=.01) and EVI (t=2.48, P=.02). Moreover, navigation efficiency exhibited distinct age-specific correlations with experienced metrics: among older adults, navigation efficiency was negatively associated with EVI (r=-0.48, P=.04), and young adults showed negative correlation between navigation efficiency and EAI (r=-0.64, P=.005).

Our findings demonstrate that k-means clustering provides an effective approach for classifying navigation states and extracting quantitative behavioral indicators for assessing cognitive abilities. In addition, the environment-based experienced metrics derived from space syntax analysis revealed distinct age-related navigation patterns, highlighting how spatial configuration shapes wayfinding behavior across age groups. These results establish an important foundation for future applications in clinical cognitive assessment and rehabilitation, as well as the design of age-friendly urban environments.