Ecological Determinants of Restorative Outcomes in Forest-Based Meditation: A Predictive Modeling Approach for Sustainable Urban Health Planning
Zhuang Shang, Lu Zhang, P. Shen, Mingli Song
Sustainability May 8, 2026 DOI: 10.3390/su18104677 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Observational study with predictive modeling Peer reviewed |
|---|---|
| Sample size | 24 |
| Population | Healthy adults assigned to one of three forest environments |
| Intervention | Forest-based meditation |
| Duration | Three repeated meditation sessions within the same condition |
| Topics | Meditation |
| Key findings | Restorative outcomes during forest-based meditation could be reliably estimated from environmental and physiological signals, with environmental variables showing stronger predictive contributions than baseline physiological indicators. The authors suggest that restorative outcomes are structured by ecological context, while noting the small sample size makes the framework a proof-of-concept rather than a fully generalizable solution. |
Abstract
Nature-based solutions (NbS) are widely considered cost-effective approaches for improving mental health in sustainable cities. However, how different ecological settings influence restorative outcomes is still not fully clear. This study examines whether mental recovery during forest-based meditation can be predicted from environmental conditions and physiological responses in real-world settings. Twenty-four healthy adults were assigned to one of three forest environments and completed three repeated meditation sessions within the same condition, yielding 72 observations. Environmental indicators (temperature, humidity, PM2.5, CO2, illuminance, wind speed, and noise) and heart rate variability (HRV) metrics were recorded. A predictive modeling framework was applied to capture nonlinear interactions between ecological exposure and physiological responses. Subject-level separation was strictly enforced to prevent data leakage. The results indicate that restorative outcomes can be reliably estimated from ecological and physiological signals under the observed conditions. Environmental variables exhibited stronger predictive contributions than baseline physiological indicators. These findings suggest that restorative outcomes are structured by ecological context. Given the limited sample size (N = 24), the proposed framework should be interpreted as a proof-of-concept model rather than a fully generalizable solution. Repeated subject-level random splits yielded consistent predictive performance across data partitions, indicating the robustness of the model.