Integrating restorative benefits into age-friendly community design: A data-driven approach using multimodal psychophysiological and visual sensing

With acceleration of population aging, it has become particularly important to promote psychophysiological restoration for older adults in community environments. Although restorative benefits are increasingly recognized as a key pathway to mental health, there is a lack of integration of restorative benefits into age-friendly community evaluation and planning. Previous studies have typically been limited by unimodal data, failing to capture complex relationships between environmental exposures and restorative responses. To address these limitations, this study develops a data-driven framework that leverages real-time visual sensing, spatial isovist analysis, and psychophysiological measurements to explore how community environments shape restorative benefits. We conducted an on-site walking experiment with 34 older adults across routes in urban and rural communities, collecting visual, spatial, and psychophysiological data. This study leveraged multilevel linear regression and generalized additive models to examine linear and nonlinear associations, and applied random forest models to identify key environmental contributors and their interaction effects. Our results showed that, beyond common natural elements such as vegetation and sky, waterbodies and soil elements had positive relationships with emotional valence and restorative evaluation (β = 0.1535 to 0.7337, p < 0.05). Visual geometric metrics, particularly isovist area, showed statistically significant associations with emotional valence, restorative evaluation and arousal level (β = −0.0156 to 0.5197, p < 0.05). Notably, community environmental factors, such as vegetation, waterbodies, and compactness, exhibited significant nonlinear and threshold relationships with psychological restoration outcomes. Finally, we identified vegetation, drift angle, and drift magnitude as dominant contributors, with strong interactions jointly influencing restorative responses. These findings emphasized that restorative responses were shaped by the combined configuration of environmental elements rather than isolated attributes. This study offers valuable insights into the restorative benefits of the visual environment, enhances Digital Twin analytics for smart community management, and has implications for the design of age-friendly communities.