Current methodologies of greenspace exposure and mental health research—a scoping review

The paper investigates how greenspace exposure relates to mental health and synthesizes the methodological approaches used across recent studies. It situates greenspace within three pathway domains (mitigation of harm, instoration through social and physical activity, and restoration), and references psychological theories such as Attention Restoration Theory (ART) and Stress Recovery Theory (SRT) to explain potential mechanisms. The work highlights that the prevalence of studies on greenspace and mental health has increased, particularly in urban contexts where mental illness risk is higher, but definitions and measurements of greenspace and mental health vary widely, hindering comparability. The review aims to answer: 1) Which greenspace and mental health methods are used in scientific studies? 2) How are different methods of greenspace exposure measurements and mental health outcomes linked? 3) What research gaps result from methodological combinations. By mapping methods and their combinations, the study seeks to inform future interdisciplinary research and practice for designing environments with mental health benefits.

Prior reviews have examined specific greenspace features (e.g., biodiversity, indoor plants, trees), particular populations (e.g., children), pandemic-related impacts, and selected methodological foci (e.g., biopsychological outcomes, GIS-based exposure). Some reviews consistently report positive associations between greenspace and mental health, while others find limited or inadequate evidence. Literature shows definitional inconsistencies of greenspace across disciplines and critiques common exposure proxies like NDVI for missing quality, type, and biodiversity detail. Mental health assessments have also varied, often relying on self-developed rather than validated tools, raising concerns about measurement quality. Existing reviews rarely provide a quantitative overview of the combined use of greenspace exposure and mental health assessment methods across studies. This review addresses that gap by cataloging and quantifying both domains and their combinations.

Design: Scoping review following PRISMA-SCR guidance. A PEO framework (Population, Exposure, Outcome) informed search terms, categorizing greenspace as exposure rather than intervention. Search: Databases—Medline (PubMed), Embase, PsycInfo, Web of Science. Search date: May 23, 2022. Languages: English or German. Years: 2017–2022. The review was not pre-registered. Inclusion criteria: Primary exposure measurement of urban/rural green exposure (direct or indirect); at least one mental health indicator; human subjects; mainly adolescents >12 and adults; empirical/experimental original studies; published 2017–2022; English/German. Exclusion criteria: Studies exclusively analyzing other exposures (e.g., air pollution); additional therapeutic intervention (e.g., counseling in greenspace); effects on animals/plants; studies mainly on children <12; published before 2017. Screening process: Two independent reviewers per stage with disagreements resolved by a third. Stages: title screening (five authors), abstract screening (two authors per paper), methodological screen focusing on methods sections. WHO-5 was excluded as assessing general well-being rather than mental health. Data extraction: Standardized charting by two authors with 10% double-charted for consistency. Extracted: publication characteristics, study population, greenspace types, greenspace exposure measurements, mental health outcomes and measurements. Categories for greenspace types, greenspace measurements, mental health outcomes, and mental health measures were iteratively synthesized with reference to prior research. No formal quality assessment performed, as per scoping review aims. PRISMA counts: Records identified (n=18,325); after duplicates (n=12,401); titles excluded (n=9,649); abstracts screened (n=2,752), abstracts excluded (n=2,333); full texts assessed (n=419), full texts excluded (n=81) for reasons including no mental health focus (n=37), publication type (n=14), additional intervention (n=12), no greenspace focus (n=12), pre-2017 (n=3), children studies (n=2), no full-text access (n=1). Final included studies: n=338.

- Study geography: China (n=66) largest share; United States (n=46); Australia (n=19). - Study designs: Cross-sectional (n=141); Experimental (n=118); Longitudinal (n=54); Qualitative (n=15); Mixed methods (n=6). - Populations: General adult population (n=213); Young adults/students (n=65); Older adults >60 (n=23); Adolescents (n=21); Patients with medical conditions (n=12); Administrative units (n=7). Some studies included multiple populations. Greenspace types (three scales): - Natural elements and species (total n=20): Trees (10), Indoor plants (6), Grasses (3), Ornamental plants (3), Greening of buildings (2), Shrubs (1). 35% virtual exposures in this scale. - Land-use types and ecosystems (total n=197): Public parks (92), Forest/woodland (66), Gardens (32), Tree cover (26), Grassland/meadow (19), Roadside greenery (11), Shrubland (9), Arable land (7), Courtyard (2), Informal greenspace (1), Cemetery (1). 18% virtual. - Landscapes including patches and non-specific greenspace (total n=184): Urban green (162), Rural greenspace (40). 9% virtual. Greenspace exposure measurements (n counts exceed studies due to multiple methods per study): - Geoinformation systems (GIS): n=145 (most common), including NDVI and land-use/land-cover metrics (abundance/proximity). - Predefined through intervention: n=120 (controlled activities, viewing scenes; images/videos/VR). - Self-reported quantitative: n=56 (questionnaires on frequency/duration/quality/proximity/usage). - Expert assessment: n=21 (researcher/expert classification or quality assessment). - Self-reported qualitative: n=18 (interviews, thematic writing, participatory methods). - Street view (SV): n=8 (eye-level imagery around residence/neighborhood/routes). - Ecological momentary assessment (EMA): n=7 (momentary exposure via self-report/GPS+GIS/SV with concurrent mental health measure). Mental health outcomes (n counts exceed studies due to multiple outcomes per study): - Affect & mood: n=148 (e.g., PANAS, STAI, POMS). - Stress: n=96 (e.g., Perceived Stress Scale; cortisol and other physiological indicators). - Mental disorder: n=90 (registry data, prescription rates; diagnostic instruments like ADAS). - Restoration: n=58 (e.g., Perceived Restorativeness Scale). - Mental health (general): n=57 (e.g., GHQ-12, SF-12). - Cognitive outcomes: n=43 (e.g., memory and attention tests). - Brain activity & structure: n=33 (MRI, fMRI, EEG). - Mental well-being: n=32 (eudaimonic/hedonic; e.g., WEMWBS). - Vitality: n=12 (Subjective Vitality scales). - Miscellaneous: n=23 (sleepiness, smoking, emotional eating). Mental health measurement methods (n counts exceed studies): - Questionnaire: n=255 (validated and author-constructed; predominantly self-report). - Physiological marker: n=62 (blood pressure, HRV, salivary cortisol, eye-tracking). - Cognitive testing: n=39 (Stroop, SART, MMSE, etc.). - Neurological indicator: n=29 (fMRI, MRI, EEG). - Epidemiological measurement: n=27 (incidence, prescription rates, length of stay; used for mental disorders only). - Qualitative measurement: n=22 (interviews, focus groups, photovoice). - Behavior or facial expression: n=8 (behavioral screening; facial emotion recognition). Combinations of methods: - Total combinations identified: n=484. - Most frequent combinations: • Predefined greenspace + Questionnaire: n=102 (highest single combination). • Predefined greenspace + Physiological markers: n=54 (most common pairing with physiological markers). • Predefined greenspace + Neurological indicators: n=24; + Cognitive testing: n=21; + Qualitative measurement: n=5; + Behavior/facial expression: n=4. • GIS + Questionnaire: n=98 (widely used, cost-effective for large populations/longitudinal designs). • GIS + Epidemiological measurement: n=27; GIS + Cognitive testing: n=17; GIS + Neurological indicators: n=4; GIS + Physiological markers: n=5; GIS + Behavior/facial expression: n=3. No GIS + Qualitative measurement reported. • Self-reported quantitative + Questionnaire: n=53; + Physiological markers: n=4; + Qualitative measurement: n=2; + Cognitive testing: n=1. • Expert assessment + Questionnaire: n=16; + Qualitative measurement: n=5; + Physiological markers: ≤2; + Epidemiological measurement: ≤2; + Behavior/facial expression: ≤2. • Self-reported qualitative + Qualitative measurement: n=17; + Questionnaire: n=2; + Physiological markers: n=1. • Street view + Questionnaire: n=7; + Epidemiological measurement: n=1. • EMA + Questionnaire: n=7; + Neurological indicator: n=1. Intervention studies—real vs virtual greenspace (predefined exposures): - Predefined exposure studies roughly split: Real greenspace n=103; Virtual greenspace n=102; both real and virtual n=4. - Within predefined exposure combinations (total methods used n=209): Questionnaire (real=49; virtual=50; both=3; total=102); Physiological markers (real=26; virtual=26; both=1; total=53); Neurological indicators (real=13; virtual=11; total=24); Cognitive testing (real=10; virtual=11; total=21); Qualitative measurement (real=4; virtual=1; total=5); Behavior/face recognition (virtual=4; total=4).

Findings reveal distinct methodological patterns used to investigate greenspace and mental health. Common pairings like GIS with questionnaires and predefined exposures with questionnaires or physiological/neurological measures indicate cost-efficiency and feasibility for large-scale or controlled studies, reinforcing evidence for certain pathways (e.g., proximity/abundance of greenspace and self-reported outcomes; immediate physiological responses to controlled exposures). However, reliance on these combinations risks imbalanced evidence and overlooks facets such as greenspace quality, perceptions, and everyday dynamic exposures. Less-used methods (e.g., expert assessments, qualitative exposure measures) can capture quality, perception, and social dimensions, while EMA and SV can better represent real-world time-activity patterns and visibility at eye level. The review also highlights inconsistent and imprecise greenspace terminology and insufficient description of virtual stimuli, which impedes comparability and synthesis. Neuroscientific indicators add mechanistic insights but currently lack diagnostic specificity for mental disorders. Overall, diversifying methodological combinations, improving reporting and definitions, and integrating interdisciplinary approaches can advance a comprehensive understanding of how different greenspaces confer mental health benefits.

This scoping review provides the first quantitative overview of the diverse methods used to assess greenspace exposure and mental health and their combinations. Many studies employ similar pairings (e.g., GIS + questionnaires; predefined exposure + questionnaires/physiological markers), strengthening evidence for selected pathways but limiting breadth. Future research should: expand mixed-methods designs combining qualitative insights with quantitative and spatial/temporal exposure measures (e.g., interviews with GPS/EMA/SV); explore underused combinations (e.g., SV with physiological markers); systematically assess greenspace quality and type; compare real versus virtual greenspaces to clarify use-cases for technologies like VR; and adopt consistent, literature-informed definitions and reporting frameworks (e.g., PRIGSHARE). Interdisciplinary collaboration across ecology, public health, psychology, medicine, and urban planning is essential to develop common terminology and robust, multifaceted evidence to guide the design of greenspaces that optimize mental health outcomes.

- Timeframe restriction (2017–2022) and language limits (English/German) may have led to omission of relevant studies. - Possible underrepresentation of rural greenspace due to search strategy focus. - Inconsistent greenspace terminology across studies could bias search, screening, and manual extraction. - Overlap among categories (e.g., predefined exposure often includes expert site selection) may reduce counts in some categories. - No formal study quality assessment performed, consistent with scoping review aims. - Review not pre-registered. - Some included studies inadvertently assessed blue spaces due to inadequate reporting and could not be excluded at eligibility.