Differential Efficacy of AI-Enhanced Virtual Nature on Mental Health Among Older Adults in Indoor Environments: A Randomized Comparison of Television, Virtual Reality, and Mixed Reality

作者：Ying Cao, Ruirui Cai, Hanbin Shen, Yuying Peng, Kexin Yu, Peiying Yu, Di Zhou · 发表于：International Journal of Human-Computer Interaction · 年份：2026 · DOI：10.1080/10447318.2026.2639124 · 被引用次数：2 · 研究领域：Virtual Reality Applications and Impacts、Technology Use by Older Adults、Urban Green Space and Health

This study develops an artificial intelligence-augmented virtual nature intervention to address the limited access to natural elements in indoor environments for older adults. Using artificial intelligence-generated natural materials, this study created virtual reality and mixed reality nature‑based scenes. A randomized controlled trial enrolled 114 elderly participants aged 60–85 years without severe cognitive or sensory impairments, who were randomly assigned to an AI-enhanced television group (n = 38), a VR group (n = 38), or a MR group (n = 38). Research assessed psychological and physiological indicators including perceived stress, positive and negative emotions, connection to nature, heart rate, and skin conductance. The randomized controlled trial results demonstrated that the VR group showed the most significant reduction in stress levels (F = 10.20, p < 0.001, η2 = 0.271). The MR group exhibited the best performance in enhancing positive emotions (F = 6.78, p = 0.002, η2 = 0.198) and increasing natural connection (F = 6.57, p = 0.003, η2 = 0.193). No adverse effects were reported.