Nature images are more visually engaging than urban images: evidence from neural oscillations in the brain

Environmental neuroscience investigates how physical environments influence cognition and brain function, with growing evidence that natural environments benefit stress, mood, and attention. ART posits that urban environments impose high attentional demands that fatigue limited resources, whereas natural environments engage involuntary, effortless attention (‘soft fascination’) allowing recovery. EEG, with millisecond temporal resolution, is well-suited to capture rapid attentional processes. This study asks whether viewing nature images, compared to urban images, modulates neural oscillations indexing visual engagement (parietal alpha) and cognitive demand (frontal theta). The authors hypothesized: H1) nature viewing would reduce parietal alpha power vs. urban viewing (greater visual engagement); H2) nature viewing would reduce frontal theta power vs. urban viewing (lower cognitive demand); H3) nature viewing would increase perceived restoration vs. urban viewing.

Prior work using fMRI and fNIRS has shown differential hemodynamic responses to natural versus urban stimuli (e.g., walking in nature; viewing images and videos; listening to nature soundscapes), but these modalities measure indirect, slower hemodynamic signals, limiting insights into rapid attentional fluctuations. EEG provides direct, temporally precise measurement of neural activity through ERPs and resting oscillations. ART’s ‘soft fascination’ suggests natural stimuli are visually engaging but not demanding. Parietal/occipital alpha power (8–12 Hz) inversely indexes visual engagement, reflecting inhibition of visual cortex excitability; reduced parietal alpha indicates greater visual attention to stimuli. Prior environmental EEG studies report mixed alpha findings, often measuring alpha at non-parietal sites; Hopman et al. (2020) specifically observed decreased parietal alpha during real nature immersion, suggesting increased visual engagement. Frontal theta (4–8 Hz), linked to dACC and effortful attention, increases with cognitive demand; higher resting frontal theta relates to poorer task performance, implying elevated cognitive load. Few studies have examined frontal theta in nature contexts; McDonnell and Strayer (2024b) found higher frontal theta after urban vs. nature walks, suggesting greater attentional demand in urban contexts. The present study adopts an a priori focus on parietal alpha and frontal theta during viewing of nature vs. urban images to align with ART.

Design: Between-subjects randomized design with two conditions: viewing nature images or urban images for 10 minutes during continuous EEG recording. Participants: N=61 (final EEG N=58 after data loss; 30 nature, 28 urban). Ages 18–44 (M=24.69, SD=5.48); 74% female, 26% male; racial/ethnic distribution reported; compensation via research credits or $20. A priori power analysis (d=0.40, power=0.80) suggested N=52 minimum; data collected exceeded this to account for attrition. Procedure: Consent; EEG electrode setup; passive viewing of a 10-minute slideshow of images; subsequent cognitive tasks (reported elsewhere); PRS short version; total session ~2 hours. Stimuli: Validated image sets (Berman et al., 2008): 50 nature images (water, vegetation, natural elements, no human-made structures) or 50 urban images (buildings, vehicles, roads, infrastructure). Each image presented for 7 s in randomized order on a loop for 10 minutes on a 13-inch laptop at 24 inches viewing distance; passive viewing instructions. EEG recording: BIOPAC BioNomadix wireless system; passive Ag/AgCl electrodes at Fz, Cz, Pz; ground forehead; reference right mastoid; VEOG with two electrodes above/below right eye; 10–20 system; impedances <10 kΩ. Signals amplified via BioNomadix Smart Center, max sampling rate 2 kHz/channel; monitored with AcqKnowledge v5.0. EEG processing: Offline processing in MATLAB using EEGLAB and ERPLAB. Downsample to 250 Hz; band-pass 0.1–30 Hz (Butterworth, 12 dB/oct). Ocular artifact correction via Gratton EMCP; moving-window artifact rejection (flatlines or peak-to-peak >200 μV). Mean data loss 0.73% overall (0.34% nature; 1.10% urban). Artifact-free data epoched into 1-s intervals with Hanning window; FFT computed to obtain power 1–30 Hz. Measures: Parietal alpha power defined as average 8–12 Hz power at Pz; frontal theta power defined as average 4–8 Hz power at Cz. Electrode choices were a priori based on prior literature and confirmed post hoc as maximal sites. Self-report: Short PRS (5 items; 1–10 scale), averaged to total score. Statistics: Assumption checks (Shapiro–Wilk for normality; Levene’s for equal variances). When violated, log transformation applied before re-testing assumptions. Independent samples t-tests on (transformed) data. Effect sizes: Cohen’s d (small=0.2, medium=0.5, large≥0.8). Visualization via matplotlib; analyses using SciPy and NumPy.

Final EEG sample N=58 (30 nature, 28 urban); final PRS sample N=58 (28 nature, 30 urban). Parietal alpha power (8–12 Hz at Pz): Nature M=2.37, SE=0.38; Urban M=4.36, SE=0.70. After log transformation, independent samples t-test showed significantly lower parietal alpha during nature viewing vs. urban viewing, t(56) = −2.34, p = 0.023, Cohen’s d = −0.62 (medium to large), indicating greater visual engagement for nature scenes. Frontal theta power (4–8 Hz at Cz): Nature M=2.94, SE=0.27; Urban M=4.30, SE=0.65. After log transformation, no significant difference, t(56) = −1.55, p = 0.126, Cohen’s d = −0.41 (small to medium), suggesting a non-significant trend toward lower cognitive demand during nature viewing. Perceived Restorativeness Scale: Nature M=7.41, SE=0.26; Urban M=3.91, SE=0.26; assumptions met; t(56) = 9.58, p < 0.001, 95% CI [2.76, 4.23], Cohen’s d = 2.51 (very large), indicating strong subjective restoration for nature images.

Findings support ART’s proposition that nature engages ‘soft fascination’: nature images led to significantly lower parietal alpha power, indexing greater visual engagement, without a concomitant significant increase in frontal theta (cognitive demand). The numerical trend toward lower frontal theta during nature viewing aligns with prior reports of reduced cognitive load in natural contexts, though replication with greater power is warranted. Results are consistent with Hopman et al. (2020) for parietal alpha decreases during nature immersion and with broader literature showing perceptual and cognitive benefits from natural stimuli. The Perceptual Fluency Account complements ART by suggesting that organic, fractal-like features of nature scenes afford easier processing, enhancing positive affect and reducing cognitive load. Together, neural and self-report data indicate that even 2D nature representations can evoke restoration by engaging attention efficiently and reducing demand, informing environmental design and interventions aimed at mitigating attentional fatigue in urbanized settings.

The study demonstrates that viewing nature images increases visual engagement (lower parietal alpha) without significantly elevating cognitive demand (frontal theta), while robustly enhancing perceived restoration. These results align with ART and the Perceptual Fluency Account, suggesting that natural scenes facilitate attentional recovery through effortless engagement. Contributions include a hypothesis-driven focus on specific oscillatory markers at theoretically relevant scalp sites during stimulus exposure, bridging neural mechanisms and subjective restoration. Future research should: 1) systematically manipulate low-level visual features (e.g., fractal complexity, edge density, spatial frequency) to isolate contributors to neural changes; 2) use normed image sets for affective and perceptual properties; 3) examine multisensory nature exposure (e.g., auditory, olfactory) and real-world, mobile neuroimaging; 4) vary exposure type and duration (static images vs. videos vs. VR; brief vs. prolonged) to identify optimal parameters; and 5) investigate individual differences (e.g., connectedness to nature, environmental experience) and potential interactions with stress recovery processes.

Low-level visual features (fractal patterns, spatial frequency, hue, edge density) were not systematically controlled or isolated, limiting attribution of specific scene properties to oscillatory changes. Image sets, chosen for consistency with prior work, were not normed on arousal, valence, mystery, or preference, which may influence engagement and demand. The focus on visual stimuli alone constrains conclusions about multisensory contributions to restoration. Reverse inference limitations apply to interpreting oscillations as direct markers of visual engagement and cognitive demand; other psychological processes may contribute. Sample size was adequate for primary effects but underpowered for investigating individual differences (e.g., connectedness to nature, familiarity with environments). Potential stress-related mechanisms (per Stress Recovery Theory) were not measured and may interact with attentional restoration.