The influence of a walk in nature on human resting brain activity: a randomized controlled trial

The study investigates how immersion in natural versus urban outdoor environments influences affect and resting neural activity associated with executive attention. Grounded in Attention Restoration Theory (ART), the authors posit that urban settings impose sustained demands on executive attention due to competing stimuli, leading to mental fatigue, whereas natural environments gently engage attention, allowing executive resources to rest and replenish. Prior behavioral work shows improved executive attention and mood after nature exposure, but laboratory task requirements may disrupt restoration processes. To address this, the authors examine a resting-state neural marker of executive attention—frontal midline theta (fMθ; 4–8 Hz), thought to originate from dACC networks—before and after a controlled 40-minute walk in nature or in an urban environment. The primary hypothesis is that a nature walk will place lower demands on executive attention, reflected in lower post-walk resting fMθ compared to an urban walk, alongside greater improvements in positive affect.

The paper situates the work within environmental neuroscience, noting global urbanization and associated health and cognitive risks (e.g., higher depression, anxiety, and attention disorders). Extensive literature documents benefits of nature exposure on executive attention, mood, creativity, stress recovery, immune function, blood pressure, and social connectedness. ART proposes that nature restores attentional capacity by providing softly fascinating stimuli that require little executive control, whereas urban environments strain directed attention. Traditional paradigms often assess task performance (e.g., Flanker, ANT) pre/post exposure, potentially confounding restoration with task demands. Neurocognitively, fMθ is tightly linked to cognitive control and increases with task difficulty and working memory load; it reflects dACC–prefrontal circuitry associated with attention allocation. The authors therefore target resting fMθ to capture spontaneous neural signatures of attentional demand/restoration without task interference.

Design: Single-blind randomized controlled trial comparing a 40-minute low-intensity nature walk versus an urban walk matched on time, distance, and elevation. Participants: N=92 (71 female, 20 male, 1 non-binary; 80% White, 10% Asian, 7% Latinx/Hispanic, 3% Black), ages 18–57 (M=29.43, SD=10.52), recruited from University of Utah and the broader Salt Lake City community. Inclusion required normal/corrected vision, ≥6 hours sleep prior night, no neurological conditions, and ability to walk comfortably for 40 minutes (~20 min/mile). Compensation: course credit or $50. Power analysis (PANGEA) targeted 86; 92 were enrolled to offset potential loss. Procedure: Upon arrival, EEG setup and demographics; PANAS for baseline affect; 2-min baseline resting EEG (eyes-open fixation). Depletion manipulation: 10-min counting backward from 1000 by 7s (out loud) to induce cognitive load; manipulation-check items (tiredness, effort, pleasantness, frustration) showed no condition differences. Then a second 2-min resting EEG and PANAS (pre-walk). Participants were randomized to: (a) Nature walk at Red Butte Garden and Arboretum or (b) Urban walk on the adjacent medical campus. Walks (~2 miles, ~40 minutes) departed from the same lab but in opposite directions. Post-walk: PANAS and a third 2-min resting EEG (post-walk). Equipment and controls: Garmin Forerunner 45S tracked GPS, time, pace, elevation, heart rate, calories; GoPro verified route adherence. Environmental variables (temperature, wind, humidity) recorded. Exercise/environmental matching: Groups were similar on distance, moving time, pace (low-intensity), average/max heart rate, calories, ascent/descent, temperature, humidity, and wind speed (Table 1). EEG acquisition: 32-channel BrainVision actiCap (10–20 system), impedances <25 kΩ; actiCHamp Plus amplifier; 500 Hz sampling; TP10 reference online; VEOG recorded with auxiliary electrodes. Resting recordings were 2-min eyes-open with fixation. EEG also recorded during depletion (as manipulation check for increased fMθ with workload). EEG preprocessing: MATLAB with EEGLAB/ERPLAB; downsampled to 250 Hz; 0.1–30 Hz Butterworth bandpass (12 dB/octave); re-referenced to average mastoids (TP9/TP10); ocular artifact correction via Gratton EMCP; artifact rejection with moving window (flatlines/peak-to-peak >200 µV). Mean data loss: 2.08% baseline, 12.40% depletion, 2.19% pre-walk, 2.37% post-walk. Artifact-free continuous data segmented into 1-s Hanning-tapered epochs. FFT derived power spectra 1–30 Hz; fMθ quantified as average power 4–8 Hz across frontocentral electrodes (Fp1, F3, Fz, F4, FC1, FC2, C3, Cz, C4, CP1, CP2, Fp2). Samples for analysis: Affect analyses used 92 at baseline and pre-walk (46/46), and 91 post-walk (45 nature, 46 urban). EEG analyses used 91 at baseline and pre-walk (45 nature, 46 urban) and 90 post-walk (44 nature, 46 urban) due to one rain interruption and one poor-quality EEG. Statistics: R 4.0.2; linear mixed-effects models (lme4) with Participant ID random intercept. For each outcome (positive affect, negative affect, fMθ), models tested fixed effects of Condition (nature, urban), Time (baseline, pre-walk, post-walk; fMθ additionally included depletion), and their interaction. Inference via likelihood ratio tests (afex::mixed). Post-hoc pairwise contrasts on estimated marginal means with Tukey adjustment (emmeans). Scalp maps plotted with MNE.

Affect (PANAS): - Positive affect showed a significant main effect of time (χ²(2)=122.95, p<0.001) and a significant Time × Condition interaction (χ²(2)=9.11, p<0.05), with no main effect of condition (χ²(1)=2.06, p=0.152). Both groups decreased from baseline to pre-walk and increased from pre- to post-walk, but the nature group had a significantly larger post-walk increase than the urban group. Post-walk nature vs urban difference estimate: 4.48 points, t(133)=2.59, p<0.05. Within-group pre–post: nature −9.02, t(179)=−9.86, p<0.001; urban −5.74, t(179)=−6.32, p<0.001. - Negative affect showed a significant main effect of time (χ²(2)=24.77, p<0.001), with no main effect of condition (χ²(1)=2.51, p=0.113) and no interaction (χ²(2)=0.32, p=0.850). Negative affect decreased from pre- to post-walk and was lower post-walk than baseline for both groups. EEG (fMθ power, µV²/Hz): - Significant main effect of time (χ²(3)=235.52, p<0.001): fMθ increased from baseline to depletion (large increase) and decreased from depletion to pre-walk but remained elevated above baseline at pre-walk (Baseline–Pre-walk difference estimate = −0.36, z=−3.60, p<0.01), confirming successful fatigue induction and residual effects of cognitive demand. - No main effect of condition (χ²(1)=0.72, p=0.396), but a significant Time × Condition interaction (χ²(3)=18.94, p<0.001). From pre- to post-walk, the urban group showed a significant increase (pre–post difference estimate = −0.65, z=−4.65, p<0.001; negative indicates post > pre), while the nature group did not (0.17, z=1.18, p=0.643). Post-walk nature vs urban difference: −0.75, z=−2.41, p<0.05 (urban higher). Exploratory: Changes in affect (positive or negative) were not significantly correlated with changes in fMθ (r(88)=−0.05, p=0.662; r(88)=0.07, p=0.535). Effects persisted after covarying age, ambient temperature, and distance walked.

Findings support ART: urban walks imposed higher executive attention demands, reflected by elevated post-walk resting fMθ, whereas nature walks allowed executive attention to rest, yielding lower post-walk fMθ. Both groups improved mood, likely due to exercise and relief from depletion, but the nature group’s greater increase in positive affect suggests additional environment-specific benefits (e.g., aesthetic qualities, a stronger sense of “being away”). The manipulation checks and fMθ increase during the depletion task validate fMθ as a neural index of cognitive effort. Elevated pre-walk fMθ versus baseline indicates residual cognitive load from the depletion task, underscoring that attentionally demanding activities leave a lingering neural signature even at rest. Critically, the fMθ differences cannot be attributed to exercise metrics or environmental conditions since groups were matched on heart rate, calories, pace, distance, elevation, temperature, humidity, and wind, and effects remained after controlling for age, temperature, and distance. The lack of correlation between affect and fMθ changes suggests partially independent mechanisms for mood improvement and attentional restoration. Collectively, results identify resting fMθ as an emergent neural signature differentiating attentional load in natural versus urban environments, advancing ecological validity by avoiding task-induced confounds.

In a single-blind randomized controlled trial (N=92), a 40-minute nature walk produced a larger increase in positive affect than a matched urban walk and did not elevate resting frontal midline theta power, whereas the urban walk increased post-walk fMθ. These results indicate greater attentional strain in urban environments and attentional rest in nature, with resting fMθ serving as a robust neural marker of environmental attentional demand. The study advances mechanistic understanding of nature-based cognitive restoration using resting-state EEG without task interference and under carefully matched physical and environmental conditions. Future work should examine other oscillatory bands and cross-frequency interactions, leverage mobile neuroimaging in situ, and include more diverse populations to improve generalizability.

- Potential unmeasured confounds remain: stress was not directly measured or manipulated; additional influences on fMθ beyond attentional demand may exist. - Demographic composition skewed predominantly White, educated, and middle-aged, limiting generalizability; broader sampling across race, SES, and age is needed. - Pre–post laboratory design; real-time mobile EEG during walking was not implemented (EEG is sensitive to movement artifacts), which may limit ecological validity despite efforts to minimize confounds. - While walks were matched on numerous metrics, minor differences in environmental features inherent to settings (e.g., unpredictable urban stimuli) cannot be fully standardized. - One participant lost to rain-related interruption; one EEG dataset excluded for quality, though overall power remained adequate.