How nature nurtures: Amygdala activity decreases as the result of a one-hour walk in nature

Urbanization is increasing worldwide and is associated with elevated risk for mental health problems including anxiety disorders, mood disorders, major depression, and schizophrenia. Urban upbringing is a strong predictor of schizophrenia incidence, potentially via increased social stress. In contrast, exposure to nature is theorized to promote restoration and stress relief, supported by frameworks such as Attention Restoration Theory (ART) and Stress Recovery Theory (SRT). Empirical work shows that nature exposure can improve cognition and reduce stress, both psychologically and physiologically (e.g., heart rate, blood pressure, cortisol). Neurobiologically, cross-sectional work indicates greater amygdala activation to social stress in urban vs. rural dwellers. One prior fMRI intervention showed a 90-minute nature walk reduced subgenual prefrontal cortex activity associated with rumination, without change after an urban walk. However, causal effects of nature vs. urban exposure on stress-related brain regions remain unclear, and it is unknown whether observed stress relief reflects benefits of nature per se versus absence of urban stressors. The present fMRI intervention study tested whether a one-hour walk in a natural environment versus an urban environment differentially alters activation in stress-related brain regions (amygdala, ACC, dlPFC) during a fearful faces task and a social stress task. The hypothesis was that stress-related regions would show reduced activation after nature exposure relative to urban exposure, compared to pre-walk baseline.

The paper situates its question within two main theoretical frameworks: Attention Restoration Theory (ART), which posits that natural environments restore directed attention through effortless fascination, and Stress Recovery Theory (SRT), which emphasizes affective recovery from stress via increased positive emotions and decreased arousal and negative affect. Empirical literature documents benefits of nature exposure on cognition (e.g., working memory, directed attention) and affect/stress (reductions in negative emotion, heart rate, blood pressure, cortisol). Neuroimaging literature shows higher amygdala reactivity to social stress in urban vs. rural residents, suggesting neural pathways for urbanicity-related risk. One fMRI intervention previously found reduced subgenual prefrontal cortex activity after a 90-minute nature walk, linked to rumination, but did not examine stress-related regions like the amygdala nor directly compare nature’s benefits to potential urban detriments. This study addresses these gaps by focusing on stress-related ROIs and testing causal effects of a one-hour environmental exposure.

Design: Randomized intervention with pre-post fMRI. Participants completed baseline questionnaires, a working memory task, and fMRI tasks, then were randomly assigned to a 60-minute walk in a natural (urban forest, Grunewald in Berlin) or urban environment (busy shopping street, Schloßstraße), followed by a post-walk fMRI session. The order of fMRI tasks was counterbalanced across subjects and held constant within subjects across sessions. Participants: N=63 healthy, right-handed, German-speaking adults (29 females; mean age 27.21 years, SD 6.61; age 18–49). Exclusion of neurological/psychiatric disorders. Randomization balanced sex across conditions. Ethics approval obtained; informed consent provided; participants compensated. Environmental exposure: 60-minute solo walk along a predefined route displayed on a map; GPS logging ensured route adherence. Participants wore an Empatica E4 wristband to record electrodermal activity, heart rate, and HRV. They were instructed not to use phones or enter shops. After 30 minutes an alarm cued turning back to starting point. Transport to/from lab by taxi. After the walk, the same fMRI tasks were repeated; an additional social-evaluative threat task (SET) was administered but is not the focus of reported fMRI results here. fMRI tasks: (1) Fearful Faces Task (FFT) to engage amygdala responses: 30 identities (15 male, 15 female) showing fearful or neutral expressions, with masked and unmasked presentations (17 ms stimuli; inter-stimulus intervals; block structure totaling 72 face presentations across 12 blocks; 8 min 28 s). Attention to fixation cross monitored via occasional color change requiring button press. Stimuli drawn from the validated FACES database. (2) Montreal Imaging Stress Task (MIST): mental arithmetic under time pressure with performance feedback relative to a (fake) better group average to induce social stress, organized in Experimental, Control, and Rest blocks. MRI acquisition: Siemens 3T (32-channel head coil). T1 MPRAGE: TR=2500 ms, TE=4.77 ms, TI=1100 ms, matrix 256×256×192, flip angle=7°, 1 mm isotropic. Functional EPI (BOLD): TR=2000 ms, TE=30 ms, flip angle=80°, 36 axial slices, slice thickness 3.0 mm, gap 20%, FOV 216 mm, voxel size 3×3×3 mm³, GRAPPA acceleration. Preprocessing: SPM12. Slice timing correction, motion correction, normalization to MNI space via unified segmentation, spatial smoothing with 6-mm FWHM Gaussian kernel. First-level modeling: FFT modeled event-related with conditions: un/masked Fear, un/masked Neutral, and response events; high-pass filter 128 s. MIST modeled block-wise (Experimental, Control, Rest); high-pass filter 520 s. ROI definition and extraction: A priori ROIs: bilateral amygdala (also subregions via SPM Anatomy Toolbox), ACC, dlPFC based on Automated Anatomical Labeling. For FFT, mean BOLD was extracted from ROIs (time window 4–6 s post-onset). Analyses considered Fear and Neutral conditions separately and pooled due to similar patterns; masked and unmasked averaged. Statistical analysis: FFT ROI analyses used two-way mixed ANOVAs with Environment (urban vs. natural; between) × Time (pre vs. post; within). Primary focus on Environment×Time interaction. Post-hoc paired t-tests within each environment assessed pre-post change. Amygdala lateralization and subregional analyses repeated with the same ANOVA. MIST whole-brain flexible factorial contrasts (Environment×Time) used FWE correction (P<0.05); no significant clusters survived. ROI analyses for MIST extracted beta values for Experimental>Rest and Control>Rest contrasts with mixed ANOVAs analogous to FFT. Behavioral and physiological (EDA, HR, HRV) data analyzed separately (reported in Supplementary Information).

• Primary ROI (FFT): Significant Environment×Time interactions in bilateral amygdala activity: • Fear condition: F(1,61)=6.11, P=0.016, η²p=0.04. • Neutral condition: F(1,61)=4.18, P=0.031, η²p=0.03. • Pooled Fear+Neutral: F(1,61)=5.81, P=0.020, η²p=0.04.
• Post-hoc paired t-tests (pooled Fear+Neutral): • Urban group: amygdala activity stable pre to post, t(30)=−0.67, P=0.506. • Nature group: significant decrease pre to post, t(31)=2.62, P=0.014. • Nature group by condition: Fear t(31)=2.77, P=0.009; Neutral t(31)=2.37, P=0.024.
• Lateralization: Interaction driven mainly by right amygdala, F(1,61)=7.00, P=0.011.
• Amygdala subregions: • Basolateral amygdala: Environment×Time F(1,61)=5.17, P=0.026; pre-post decrease after nature walk marginal, t(31)=1.98, P=0.057. • Central amygdala: Environment×Time F(1,61)=4.31, P=0.043.
• ACC and dlPFC: No significant Environment×Time interactions in FFT (Fear, Neutral, or pooled).
• MIST: • Whole-brain: No significant clusters for Environment×Time at FWE P<0.05. • ROI patterns consistent with FFT for amygdala (descriptive decrease after nature; stable after urban); no interactions in ACC or dlPFC.
• Behavioral/physiological: • No significant Environment×Time interactions for self-report measures tied to cognitive tasks or for physiological stress indicators (EDA, HR, HRV). • Perceived restorativeness higher after nature vs. urban walk: Z=−3.85, P<0.001, r=0.49. • Enjoyment higher after nature walk: Z=−2.87, P=0.004, r=0.37. Overall: A one-hour nature walk decreased amygdala activation during affective face processing (masked and unmasked; fearful and neutral) relative to an urban walk, which showed no change. Effects localized primarily to the amygdala (not ACC/dlPFC) and were stronger in the right amygdala.

Findings support a causal, salutogenic effect of acute nature exposure on stress-related neural mechanisms. Specifically, amygdala activation decreased after a one-hour walk in a forest environment, whereas it remained unchanged after urban exposure, indicating restoration rather than urban-induced exacerbation. The decrease was evident for both fearful and neutral faces and for masked stimuli, suggesting effects can occur without conscious awareness. Right-lateralized and subregional (basolateral, central) amygdala analyses align with literature on threat processing and conditioning. Similar patterns during the social stress task (MIST) suggest generalizability to stress contexts. Lack of effects in ACC and dlPFC aligns more with Stress Recovery Theory (stress reduction) than with Attention Restoration Theory (cognition-specific recovery), at least at the neural level measured here. The neural findings provide a plausible mechanism for population-level observations linking rural living and proximity to urban forests with reduced amygdala reactivity and greater amygdala structural integrity. Practically, results underscore the potential of accessible urban green spaces to mitigate stress-related neural responses among city dwellers.

This fMRI intervention study demonstrates that a single one-hour walk in nature, compared to an urban environment, leads to decreased amygdala activity—a key stress-related neural marker—while showing no change in ACC or dlPFC. The effect appears robust across masked/unmasked and fearful/neutral facial stimuli and is stronger in the right amygdala, with convergent patterns during a social stress task. These findings provide causal evidence for the salutogenic effects of nature on stress-related brain function and suggest that increasing access to urban green spaces could benefit mental health. Future research should identify which specific environmental features drive the neural benefits, assess generalizability across diverse natural settings and cultures, disentangle behavioral measurement timing, and test cumulative effects of repeated exposures.

• Masked stimuli awareness was not explicitly tested; thus, unconscious perception cannot be definitively confirmed.
• The specific features of nature driving amygdala reduction (e.g., visual greens, sounds, odors, textures) were not isolated.
• Potential confounding by social exposure: participants in the forest may have encountered others, possibly contributing to relaxation differences versus the urban walk.
• Effects may vary across types of natural environments (e.g., tended vs. wild forests, parks, botanical gardens); only one forest route was tested.
• Cultural differences in the meaning and appraisal of nature could modulate effects; the sample was not culturally diverse.
• Behavioral questionnaires post-walk may have primarily captured stress from the subsequent fMRI tasks rather than the walk’s immediate effects, potentially masking behavioral changes.