Enhancing Restoration in Urban Waterfront Spaces: Environmental Features, Visual Behavior, and Design Implications

Abstract

Urbanization poses mental health risks for urban dwellers, whereas natural environments offer mental health benefits by providing restorative experiences through visual stimuli. While urban waterfront spaces are recognized for their mental restorative potential, the specific environmental features and individual visual behaviors that drive these benefits remain inadequately understood. Grounded in restorative environments theory, this study investigates how these factors jointly influence restoration. Employing a controlled laboratory experiment, subjects viewed real-life images of nine representative spatial locations from the waterfront space of Guangzhou Long Bund. Data collected during the multimodal experiments included subjective scales data (SRRS), physiological measurement data (SCR; LF/HF), and eye-tracking data. Key findings revealed the following: (1) The element visibility rate and visual characteristics of plant and building elements significantly influence restorative benefits. (2) Spatial configuration attributes (degree of enclosure, spatial hierarchy, and depth perception) regulate restorative benefits. (3) Visual behavior patterns (attributes of fixation points, fixation duration, and moderate dispersion of fixations) are significantly associated with restoration benefits. These findings advance the understanding of the mechanisms linking environmental stimuli, visual behavior, and psychological restorative benefits. They translate into evidence-based design principles for urban waterfront spaces. This study provides a refined perspective and empirical foundation for enhancing the restorative benefits of urban waterfront spaces through design.

1. Introduction

Urbanization has significantly altered the urban ecological landscape and the lifestyles of urban dwellers [1,2,3], exacerbating the vulnerability of human health. The landscape features and spatial quality of the urban built environment continuously influence the mental health of individuals. For instance, sensory overload in concrete-dominated environments and the reduction in the quantity and quality of natural exposure can contribute to increased occurrences of depression, anxiety, and sustained stress among individuals [4,5,6]. It is worth noting that this vulnerability to mental health issues is particularly pronounced among young adults, who exhibit heightened sensitivity to environmental stressors and are continuously subjected to academic and occupational pressures [7,8]. Therefore, this study focuses on evidence-based spatial interventions for young adults, exploring optimized design methods to stimulate the positive impact of the environment on young adults, helping to enhance individuals’ ability to regulate stress and maintain mental health.

Restorative environments are a critical area of research in the field of mental health, with two core theories—Stress Reduction Theory (SRT) and Attention Restoration Theory (ART)—providing the theoretical foundation for this study. SRT posits that emotional responses are triggered by specific environmental stimuli [9,10], while ART suggests that involuntary attention being held without any effort expended (soft fascination) is a central component of a restorative experience [11,12,13]. Additionally, a substantial amount of objective evidence supports the restorative benefits of natural environments [14,15,16,17] and indicates that direct perceptual experiences (i.e., visual perception) in natural environments have positive effects on individuals, including promoting the renewal of directed attention, the transformation of positive emotions, and the recovery from stress states, promoting psychological restoration [18].

Waterfront spaces possess the dual natural advantages of water and plants, and their restorative potential has gradually attracted attention [19,20,21]. Urban waterfront spaces in high-density environments provide residents with critical access to blue–green spaces and unique opportunities for restoration. However, there are still two research gaps, hindering evidence-based restorative design for urban waterfront spaces:

(1)

Although existing research has extensively explored the restorative effects of semantic-level features [22,23,24,25,26,27], this approach cannot reveal all of the visual environmental features that influence environmental restorative benefits [28]. It is crucial to note that the unique environmental features of highly urbanized waterfront spaces, including the multi-element interactive features of semantic-level features and the spatial morphological and sequence dimensions of scene-level features in urban waterfront spaces, have not been sufficiently studied in terms of their differentiated effects on environmental restorative benefits. There is a lack of a systematic research framework that integrates design considerations.

(2)

In previous studies, the mechanism through which environmental features translate into restorative benefits has not been validated, i.e., there is a lack of intermediary tools to quantify visual perception processes, which hinders the understanding of how visual environmental stimuli attract attention and promote restoration—an inherent limitation in subjective and objective restorative assessments (subjective scales and physiological measures), and eye-tracking technology provides a scientific approach to overcoming this limitation [29,30,31,32].

To bridge these gaps and propose actionable insights, the core research question of this study is as follows: How do environmental features shape restorative experiences through visual behavior in highly urbanized waterfront spaces, and under what mechanisms can design interventions enhance restorative benefits?

Given these two gaps, there is an urgent need for a comprehensive approach that combines the quantification of environmental features with visual behavior analysis to reveal the psychological recovery mechanisms of urban waterfront spaces. This study employs a multimodal experimental research method, combining subjective and objective restorative assessments with eye-tracking technology, to provide more specific and accurate evidence for the restorative design of urban waterfront spaces: quantifying the restorative benefits of urban waterfront spaces and the visibility thresholds of key elements, understanding how environmental features influence restorative benefits through visual behaviors, and providing evidence-based design principles to enhance the restorative benefits of urban waterfront spaces.

This study aims to develop and validate a comprehensive research framework that links waterfront space design dimensions, restorative environmental theories, and neuroscientific methods to systematically explore the mechanisms by which environmental features and visual behavior influence restorative benefits. This study makes dual innovative contributions. Theoretically, it identifies the key environmental features that affect restorative benefits in urban waterfront spaces and reveals the important role of individual visual behavior in influencing restorative benefits during the restorative experience. Practically, based on the research findings, it summarizes three major optimization principles. From exploring mechanisms to proposing principles, this study provides a more refined research perspective and evidence-based basis for the restorative design of urban waterfront spaces.

2. Literature Review

Well-designed urban environments have been found to offer psychological restoration potential similar to that of natural environments [33]. Therefore, research on their restorative qualities has gained increasing attention. Previous studies mainly focused on the restorative benefits of green spaces (e.g., parks; green areas) [34,35,36,37], while blue spaces (e.g., rivers; lakes) are often discussed together with green spaces. Some studies have emphasized the mental health benefits of water elements in the environment. Early studies by Korpela et al. (2010) observed that waterside settings were notably more beneficial than constructed green areas, approaching the psychological benefits of urban woodlands [19]. White et al. (2010) similarly demonstrated that water had positive psychological effects in both natural and artificial environments, with the psychological benefits of built environments with water being comparable to those of green spaces [38]. Moreover, studies have demonstrated that blue spaces provide health benefits through various pathways, including perceived preference, landscape design, emotional interaction, and restorative activities [39]. From reducing fatigue, improving mood, and lowering stress levels [21,40,41,42] to alleviating depressive symptoms [43], blue spaces offer health benefits that promote mental health and prevent mental illnesses.

Based on the above evidence, urban dwellers can benefit from exposure to green and blue spaces within cities, which possess restorative qualities that can promote psychological restoration. As a scarce composite ecosystem in high-density cities, urban waterfront spaces rely on water and plants to form unique advantages in psychological restoration. However, during the functional transformation from production to living, the problem of insufficient health benefits has gradually emerged. Research shows that design interventions can stimulate the health-promoting potential of waterfront environments and enhance restorative benefits [44]. To maximize the restorative potential of urban waterfront spaces, it is essential to conduct in-depth research on how environmental features influence restorative benefits, ensuring their functionality and improving their restorative qualities.

The influence of the natural environment on psychological restoration is regulated by the perception of the natural environment [45]. Among the multi-sensory pathways for perceiving environmental stimuli, vision is the primary and dominant pathway. Research has shown that the perception of the natural environment not only stems from the processing of high-level semantic information by the visual system but also includes the drive of low-level visual features, providing a visual-level explanation for the mechanism by which the natural environment promotes mental health [22,23]. Additionally, studies have demonstrated a significant correlation between the visual features of the urban waterfront space environment and restorative benefits. Prior studies by Li et al. (2023) demonstrated that the rich spatial color hierarchy and diverse plant community morphology in the urban park waterfront space can enhance restorative benefits, while a higher proportion of building and facility elements and complex spatial composition can impede these benefits [46]. Sun et al. (2024) demonstrated that the proportion of plants and water, as well as plant diversity in urban waterfront greenway, are positively correlated with positive emotions, while the width of pathways shows a negative correlation [47]. Cao et al. (2024) demonstrated that features such as plant color diversity and the aesthetic quality of water scenery have a significant predictive effect on restorative potential [48].

Based on the above evidence, the differences in the features of visual elements in urban waterfront spaces lead to differences in restorative benefits. Therefore, it is essential to investigate how environmental features can influence the restorative benefits at the visual level and to identify key visual features that can amplify these benefits. Nevertheless, there are two limitations in current research: first, the research on the features of visual elements mostly focuses on the explicit features of plants and water, and the relationship between the visibility thresholds and visual features of elements and restorative benefits under the synergistic action of multiple elements has not been revealed; second, research on environmental features primarily focuses on semantic-level features at the visual level, with limited exploration of spatial dimensions and significant gaps in the investigation of scene-level features such as spatial morphology and landscape sequences.

In addition, previous studies have not conducted an in-depth analysis of the visual attention mechanisms involved in individual environmental perception [46,47,48,49]. While subjective scales and physiological data can assess the restorative benefits of an environment to determine whether restoration occurs and to what extent, they cannot explain which environmental features within that environment (or image) contribute to such benefits [29,31].

Therefore, understanding how environmental features contribute to restorative benefits requires investigating the key mediating processes through which environmental features are perceived and processed: visual behavior—that is, what is looked at and how it is looked at. Eye-tracking technology can accurately capture human visual behavior, facilitating a detailed examination and in-depth analysis of the factors influencing the restorative benefits at a visual level [50,51,52,53]. Prior studies by Fei et al. (2023), through analysis of eye movement behavior, demonstrated that complex landscape elements and dispersed areas of interest significantly enhance restorative benefits, while single landscape elements and concentrated attention distribution have the opposite effect [54]. Fleming et al. (2024) demonstrated that subjects who gazed at natural elements had significantly reduced anxiety levels, while focusing on artificial elements inhibited restoration [55]. The above study revealed the influence of visual attention distribution and gaze content on restorative benefits, validating the empirical advantages of eye-tracking technology in restorative environment research.

Therefore, this study introduces eye-tracking technology and designs multimodal experimental research on the environmental features and restorative benefits of urban waterfront spaces to enhance the understanding of environmental restoration mechanisms and investigate avenues for enhancing psychological restoration through design optimization.

3. Materials and Methods

3.1. Research Framework

To investigate the mechanisms by which the environmental features and individual visual behaviors influence the restorative benefits of urban waterfront spaces, this study constructed a research framework (Figure 1) and designed a multimodal experiment.

This study selected three dimensions—elements, morphology, and sequences—in urban waterfront space design to fill research gaps and promote evidence-based design. First, since urban waterfront spaces blend natural (water, plants, and sky) and built elements (buildings and ground), the element dimension needs to consider the balance of these elements (multi-element interaction) and its impact on restoration. Secondly, since waterfront spaces are constrained by the water–land interface and have linear characteristics, morphology control is required to avoid excessive openness or closure, and sequence design is used to create a continuous and rich visual experience.

According to ART, a restorative environment must have four factors to promote the restoration of attentional fatigue: being away, fascination, extent, and compatibility [12]. According to SRT, the features of a restorative environment include an orderly structure, a certain spatial depth, moderate complexity, and natural environment content dominated by plants or water [9].

In order to explore the relationship between the environmental features of urban waterfront spaces in three different dimensions and their restorative benefits, this study associates these three dimensions with the environmental features proposed by ART and SRT theories (

Table 1. Theoretical mapping of three design dimensions.

Dimensions in Urban Waterfront Space Design	ART’s Features	SRT’s Features	Environmental Features of Urban Waterfront Spaces
Elements	Fascination	Environmental content	Five visual elements visibility thresholds and their features
Morphology	Fascination and extent	Gross structural and complexity properties	Degree of enclosure and interface features
Sequences	Fascination and extent	Depth properties and complexity properties	Hierarchical structure and depth perception

) and translates abstract features into actionable design elements for urban waterfront space design. Among these, “Elements” (five visual elements visibility thresholds and their features) trigger content attraction and drive restorative experiences; “Morphology” (degree of enclosure and interface features) ensures structural safety and provides comfort and a sense of security; “Sequence” (hierarchical structure and depth perception) promotes exploratory experiences and maintains sustained attraction.

Furthermore, this study employed a multimodal experimental research method, which combines subjective scales, physiological measures, and eye-tracking technology to gain a more nuanced understanding of how the mentioned environmental features influence restorative benefits. In this study, the visibility of elements was the independent variable, restorative benefits the dependent variable, and the morphological and sequential dimensions of urban waterfront spaces were supplementary explanatory variables. Restorative benefits were evaluated through subjective scale scores and physiological measurement data, while eye-tracking technology was used to capture subjects’ eye movement data. Focus heat maps and gaze plot maps were employed to visualize their “bottom-up” visual attention patterns, thereby deepening our understanding of the “visual environmental stimuli-restorative benefits” mechanism and providing evidence for restorative design.

3.2. Study Area

This study examines the waterfront space of Guangzhou Long Bund, China, specifically focusing on the stretch between Renmin Bridge and Haiyin Bridge. The spatial boundaries of this study encompass the area between the water’s edge and the adjacent urban municipal roads.

As a core segment of Guangzhou’s Pearl River landscape belt, the waterfront space of Guangzhou Long Bund represents a potential restorative environment within the city’s densely built-up area. Firstly, this site serves as an easily accessible and frequently utilized recreational destination for residents in their daily lives. The study area is adjacent to urban roads, featuring multiple bus stops, a metro station, three boat docks, and several waterborne sightseeing routes. Field observations indicate that the site experiences continuous human activity at different times of the day, including walking, exercising, and fishing. Secondly, distinct segments of the site exhibit significant variations, encompassing diverse spatial typologies including linear waterfront greenways, parks, and squares. The excellent accessibility, frequent and regular patterns of human activity, and the diverse features of the waterfront environment collectively provide an ideal setting for this study to explore the potential restorative benefits of urban waterfront spaces and the influence of visually perceived environmental factors affecting restoration.

This study conducted field surveys, capturing 83 photographs at 50 m intervals. The visibility rates of five visual elements (water, plants, buildings, ground, and sky) in these images were quantified using image semantic segmentation and manual inspection (Figure 2).

As shown in Figure 3, the element visibility analysis diagram reveals the dynamic evolution patterns of environmental features within the study area. The main observations include the following: (1) all sections exhibit low visibility features for water elements; (2) plants, buildings, and sky elements demonstrate distinct differential distribution patterns in sections A, C, and D. Based on the visibility rate variation features, this study selected representative nodes as sampling points for experimental stimulus materials.

Based on these findings, two stimulus material groups were designed: a water visibility gradient group (n = 3) and an element-dominant group (n = 2 per subgroup). Spatial nodes representing low-, medium-, and high-water-visibility levels, as well as nodes reflecting the distinct distributions of plant, sky, and building elements (further categorized into dominant, sub-dominant, or non-dominant elements), were specifically selected. A total of nine stimulus sampling nodes were used (Figure 4), comprising the water-gradient group (W1, W2, and W3), the plant-dominated group (P1 and P2), the sky-dominated group (S1 and S2), and the building-dominated group (B1 and B2).

3.3. Subjects

This study recruited 21 young adult subjects (18–30 years) from varied professional backgrounds. The study subjects were recruited through purposeful sampling using social media groups within the university to include young adults experiencing typical academic and occupational stress. The sample comprised 27.27% males and 72.73% females. All subjects exhibited sound mental health and no visual impairments, ensuring the validity of the experimental data.

3.4. Materials

3.4.1. Visual Stimulus Materials

This study utilized a camera positioned at a height of 1.6 m along the spatial midline to capture nine types of stimulus materials (Figure 4). Using Photoshop, brightness, contrast, and color differences were standardized.

To investigate the relationship between elemental visibility thresholds and restorative benefits under controlled experimental conditions, images taken at sampling points were processed to adjust the visibility of key environmental elements (water, plant, sky, and buildings) in each stimulus image, creating a gradient level. The visibility gradient adjustment rules for elements are as follows: the water visibility gradient group is divided into three levels, including high (>25%), medium (10–25%), and low (<10%), while the element-dominant group controls the gradient levels of three key elements as dominant (>35%), sub-dominant (10–35%), or non-dominant (<10%). Nine stimulus material images were generated, with the visibility ratios for each stimulus material detailed in the Appendix A.

3.4.2. Measurement of Psychological Data

The Self-Rating Restoration Scale (SRRS) [56] was selected for subjective recovery assessment because it integrates two core theoretical frameworks—ART and SRT. It includes four dimensions (emotional, physiological, cognitive, and behavioral), eight items, and a 9-point Likert scale (Figure 5). It is more comprehensive, concise, and has good validity and reliability.

3.4.3. Measurement of Physiological Data

Physiological data from subjects were collected using Kingfar ErgoLAB EDA wireless electrodermal activity sensor and PPG wireless pulse sensor (Kingfar, Beijing, China). This study focused on two indicators: skin conductance response (SCR) and the low-frequency-to-high-frequency ratio (LF/HF). SCR serves as a quantitative measure of emotional arousal and quickly reflects an individual’s response to stimulus events. LF/HF assesses the relative activity of the sympathetic and parasympathetic nervous systems, indicating the body’s response to stimuli.

3.4.4. Measurement of Eye Movement Data

Eye Movement Data from subjects was collected using Tobii Pro Glasses 3 (Tobii, Stockholm, Sweden) and visualized as focus heat maps and gaze plot maps to reflect the spatial distribution of the subjects’ attention and the sequential distribution of fixation points.

3.5. Procedure

3.5.1. Experimental Environment

The light, temperature, and noise in the experimental environment were controlled, and the stimuli were presented on a display screen. The subjects were seated three meters from the screen and followed the experimenter’s instructions throughout the experiment.

3.5.2. Experimental Procedure

The subjects were first informed of this study’s purpose and procedure (Figure 6), after which they provided informed consent and completed a demographic questionnaire. Physiological measurement devices were then fitted on the subjects. The experiment comprised three phases: a 4-min stress-inducing task involving an impromptu speech and mathematical operations; wearing of the eye-tracker and explanation of the eye-tracking protocol; and completion of the fixation task and SRRS scale as instructed. Physiological data were continuously recorded during the experiment. After verifying successful data collection, the experimenter assisted subjects in removing the equipment, concluding the experiment.

3.6. Data Processing

The SRRS scale scores, physiological data, and eye movement data from 22 subjects were collected and processed.

First, using the data processing method [56], the four dimensions of the SRRS scale were calculated and averaged to form the total score, where higher scores indicate greater restorative benefits. Second, mean values of SCR and LF/HF during the pressure application phase and across nine stimulation periods were extracted using the ErgoLAB 3.0 Platform (Kingfar, Beijing, China). A lower mean value during stimulation compared to the pressure phase signifies greater restorative benefits. Additionally, Areas of Interest (AOIs) for five element types within the nine stimulation materials were delineated using the ErgoLAB 3.0 Platform. Total and average fixation times were exported, and focus heat maps and gaze plot maps were generated. Following data processing, 19 valid datasets were analyzed using SPSS 27.0 to derive experimental conclusions.

4. Results

4.1. Environmental Types and Restorative Benefits

4.1.1. Results of Psychological Data

This study examined differences in total scores and subdimension scores across various stimulus materials (

Table 2. One-way analysis of variance for each dimension and the total score of the SRRS scale.

	Total Scores	Emotional Dimension	Physiological Dimension	Cognitive Dimension	Behavioral Dimension
F	2.933	1.101	0.23	2.075	2.097
p	0.004 **	0.365	0.985	0.041 *	0.039 *

* p < 0.05. ** p ≤ 0.01.

). A one-way ANOVA revealed significant differences in total scores (F = 2.933, p = 0.004). Subsequent analyses at the subdimension level showed significant differences in the cognitive (F = 2.075, p = 0.041) and behavioral (F = 2.097, p = 0.039) dimensions, but no statistically significant differences in the emotional and physiological dimensions (p > 0.05). Post-hoc LSD tests further indicated that stimulus material P1 differed significantly from all other materials except P2, and P2 differed significantly from W3 and B2 (Figure 7 and Figure 8).

4.1.2. Results of Physiological Data

As shown in

Table 3. One-way analysis of variance for physiological data of SCR and LF/HF.

	SCR	LF/HF
F	0.454	1.123
p	0.887	0.350

, there were no significant differences in SCR values across the different stimulus materials (F = 0.454, p = 0.887). Analysis of the SCR data (Figure 9) revealed that stimuli P2 (0.147 ± 0.133) and B2 (0.156 ± 0.258) induced the lowest arousal, indicative of a relaxed state, while W1 (0.501 ± 1.471) triggered a higher emotional response. Furthermore, the analysis of the LF/HF data (Figure 10) showed no significant differences among the stimulus materials (F = 1.123, p = 0.350). The LF/HF data for stimulus P2 (0.599 ± 0.522) was the lowest, suggesting a reduction in physiological stress, while S2 (2.749 ± 6.999) had the highest LF/HF data, indicating a higher level of stress.

4.1.3. Ranking of Comprehensive Restorative Benefits

As presented in

Table 4. Combining the results of the assessment of subjective and objective restorative benefits.

High Restorative Benefit			Medium Restorative Benefit			Low Restorative Benefit
P2	P1	W2	B2	S1	W3	S2	W1	B1

, the nine stimuli were categorized into three groups based on the psychological and physiological data: high restoration benefit (P2, P1, and W2), medium restoration benefit (B2, W3, and S1), and low restoration benefit (S2, W1, and B1).

4.2. Environmental Features and Restorative Benefits

4.2.1. Correlation Analysis Between Element Visibility and Subjective and Objective Data

Correlation analysis indicated a significant relationship between the visibility of elements and subjective questionnaire scores. Plant visibility showed a positive correlation with the SRRS total score (r = 0.284, p < 0.001), whereas building (r = −0.213, p = 0.005) and ground visibility (r = −0.216, p = 0.005) were negatively correlated. Moreover, no significant correlation was found between visibility rates and physiological indicators (SCR, LF/HF).

4.2.2. Correlation Analysis Between Element Visibility and Fixation Time

Correlation analysis shows that the visibility of water, plants, and buildings positively correlates with total fixation time, whereas the ground and sky exhibit weaker correlations. Further examination reveals that average fixation time is weakly correlated with element visibility and is predominantly affected by the visual features of elements.

Comprehensive analysis indicates that variations in visual features are key determinants of fixation time. Water, plants, and buildings enhance visual salience due to their rich information content, such as water ripples, plant details, and building structures, which stimulate exploratory interest. Consequently, high visibility rates increase fixation frequency, and element complexity extends individual fixation duration. Conversely, the sky, as a background with limited information, fails to sustain prolonged fixation despite high visibility. Ground fixation time is primarily influenced by spatial cognitive needs, such as spatial structure and movement paths, and its visual features, rather than visibility rate alone.

4.2.3. Environmental Features and Restorative Benefits

Comprehensive evaluations of both subjective and objective restorative benefits indicate that scenarios with high natural perception (e.g., P1) yield significantly higher restoration scores than those with low natural perception (e.g., B2), underscoring the impact of natural preference on subjective assessments. However, scenarios dominated by buildings with minimal plant presence (e.g., B2) still provide restorative benefits through emotional enhancement and stress reduction, suggesting that the objective restoration process is not linearly correlated with naturalness. Thus, the restorative benefits of urban waterfront spaces are influenced by the synergistic effects of multidimensional environmental features. The following analysis explores the influencing mechanisms across three design dimensions: elements, morphologies, and sequences.

In the element dimension, within-group comparisons found that the visibility of elements and visual features significantly affects restorative benefits. Regarding water element, an optimal proportion between 10% and 20% (e.g., W2) was found to best balance hydrophilicity and openness, maximizing restorative benefits. However, due to the features of the Pearl River water, exceeding this threshold appeared to induce visual discomfort. Conversely, plant proportion demonstrated a dominant restorative effect when reaching or exceeding 50% (e.g., P1 and P2). For scenarios with lower plant proportions (e.g., S2), enhancing visual complexity and appeal, and thus restorative potential, can be achieved through hierarchical configurations involving tree-layer shading, shrub-layer enclosure, and ground-cover-layer boundary softening. Building visibility became a concern at proportions ≥20% (e.g., B1 and B2), but negative perceptions could be mitigated through strategies such as optimizing facades, designing transparent interfaces, or employing plant groups for visual occlusion. Similarly, ground coverage ≥20% (e.g., S1 and W2) warranted attention; visibility could be reduced via ecological transformation or landscape penetration design, while maintaining a minimum of 15% to ensure functional passage and activities. Enhancing the richness of this bottom interface is achievable through variations in material textures. Finally, sky proportion interacted with other elements: when paired with highly visible plants (e.g., W1 and W2), maintaining 5–15% helped alleviate oppression, whereas a proportion ≥20% enhanced openness when paired with highly visible buildings (e.g., B1 and B2).

In the morphological dimension, spatial enclosure significantly influences restorative benefits. Excessive enclosure can induce a sense of oppression, while insufficient enclosure may result in a lack of perceived security. Buildings and plants, as vertical interface elements, modulate these benefits through their height, density, and interface features. In spaces dominated by buildings, controlling spacing and interface permeability is crucial. For plant-dominated areas, a sparse and hierarchical configuration is recommended to achieve a perceptual balance between “shelter” and “openness”.

In the sequence dimension, spatial hierarchy and depth perception significantly influence restorative benefits. The three-dimensional spatial structure, comprising the foreground, midground, and background, contributes to a sense of depth and facilitates natural exploration of the space. Sight corridors naturally guide the viewer’s gaze deeper, enhancing restorative effects, while obstructed views inhibit them. Moreover, the orderly arrangement of plant elements and the coherent facade design of buildings enhance spatial order, providing a clear structure that reduces visual chaos and alleviates stress.

4.3. Eye Movement Behavior and Restorative Benefits

4.3.1. Spatial Distribution of Attention and Restorative Benefits

The focus heat map presented in

Table 5. Focus heat maps of stimulus materials.

High Restorative Benefit	Medium Restorative Benefit	Low Restorative Benefit
P2	B2	S2
P1	S1	W1
W2	W3	B1

reveals that subjects’ attention was predominantly distributed along the horizontal middle-view region, with a primary focus on the bottom interface (walkway) and vertical interfaces (vegetation, structures, and building facades). The concentration and dispersion of attention reflect the relative attractiveness of these elements. Integrating this with the restorative benefits associated with each stimulus material, it can be inferred that the element features and fixation time within the middle-view focus area influence the restorative outcomes. Specifically, elements of moderate complexity (e.g., P1 and B2) appear to balance the cognitive load through a combination of explorative details and a coherent overall spatial organization, thereby attracting attention without excessive cognitive demand. Furthermore, the balanced distribution of fixation time across focal elements and the appropriate spatial layout dispersion facilitate timely attention shifts, which in turn enhance the restorative benefits.

4.3.2. Distribution of Fixation Point Sequences and Restorative Benefits

As shown in

Table 6. Gaze plot maps of stimulus materials.

High Restorative Benefit	Medium Restorative Benefit	Low Restorative Benefit
P2	B2	S2
P1	S1	W1
W2	W3	B1

, the gaze plot map shows that elements with prominent visual features (such as clear contours and unique structures) attract attention first. Elements and their layouts can guide the subjects’ gazes to conduct an orderly exploration. Among them, the ground is the fundamental element for constructing an overall understanding of the environment, plants are the key elements that attract attention and trigger further exploration, and buildings are the fixation-transfer hubs that drive the gaze to migrate to surrounding elements.

Combining the restorative benefits of each stimulus material, the following can be known:

(1)

Being attracted first by positive visual features is beneficial for the restorative benefits;

(2)

Reasonably arranging plant and building elements to guide the gaze to transfer between elements at different distances, creating a coherent exploration sequence composed of core focal points and secondary nodes, can effectively enhance the restorative benefits.

5. Discussion

5.1. Methodological Contribution and the Integrated Framework

Previous research on environmental restoration often faces challenges in holistically capturing the interplay between multi-dimensional spatial characteristics, perceptual processes (like visual behavior), and psychological outcomes. Studies frequently rely on subjective assessments (e.g., questionnaires) or focus on isolated environmental variables, potentially overlooking the mediating role of dynamic human–environment interactions. Our framework addresses this gap by systematically linking the following: objectively quantifiable design dimensions (elements), subjective and objective restorative assessments, and crucially, objective measures of individual visual behavior (spatial distribution of attention and distribution of fixation point sequences).

The successful application of this framework, as demonstrated by our results, confirms its utility and effectiveness. This framework provides a powerful tool for advancing evidence-based design aimed at enhancing mental restoration in urban settings and holds promise for investigating restorative environments more broadly, such as urban parks, forests, or even interior spaces, where visual perception plays a key role. Additionally, research has shown that the use of eye-tracking technology enhances the comprehension of psychological recovery mechanisms. By combining this technology with subjective and objective assessment methods, researchers can overcome the limitations of previous studies, a goal that is difficult to achieve using traditional methods alone.

5.2. Effects of Environmental Features on Restorative Benefits

This research reveals that the restorative benefits of urban waterfront spaces are not simply determined by their naturalness.

In the element dimension, they are jointly influenced by the ratio thresholds and visual features of the constituent elements. An optimal ratio of “5:3:2 (plants–buildings, ground–water, sky)” can maximize the restorative potential of the space. Visual complexity can compensate for insufficient proportions of elements, and positive visual attributes of water bodies, plants, and buildings can enhance the restorative benefits. Consistent with hierarchical visual processing models [22,23,45], the findings confirm that restorative benefits stem from quantifiable low-level visual features. Crucially, the findings further reveal the synergistic influence of spatial form and sequence on restorative effects.

In the morphological dimension, a balance between “sheltered-open” spatial configurations enhances restorative effects, and the transparency of building interfaces has a stronger regulatory influence on the sense of enclosure than plant density.

In the sequence dimension, a hierarchical progression from foreground to middle-ground to background (e.g., sight corridor) can extend the visual exploration path, thereby enhancing the sense of immersion. Perceptions of order (e.g., rhythmic building facades) may alleviate the influence of negative element features by reducing cognitive load.

Although Li Jie et al. [46] observed that recovery effects are inhibited at high building ratios, research has found that when architectural elements are visually optimized (e.g., transparent interfaces; facades with a strong sense of rhythm), they can become assets with restorative benefits. This resolves previous inconsistencies by emphasizing characteristic qualities rather than element categories, shifting the focus from the ‘nature-artificial’ dichotomy to collaborative visual thresholds.

Consequently, research on restorative environments should reassess the limitations of the “natural–artificial” binary paradigm. By managing visual element thresholds and optimizing features such as elements, forms, and sequences, a balance between functionality and the restorative properties of urban waterfront spaces can be achieved, maximizing restorative benefits despite limited land use.

5.3. Pathways of the Role of Visual Behavior in Restorative Experience

Analyzing eye movement behavior reveals a link between visual patterns and restorative benefits. Moderate dispersion of visual focus promotes psychological restoration more effectively than a single strong focus or a chaotic distribution. This study positioned our eye-tracking results as a methodological and mechanistic advance beyond Fei and Fleming et al. Whereas Fei [54] linked dispersed attention to restoration generally, the eye-tracking data in this study reveal how spatial features guide this process: moderate dispersion (enabled by layered sequences) maximizes soft fascination. Designing environments with moderate complexity and hierarchical element distribution can enhance these restorative benefits. Unlike Fleming et al. [55] (who contrasted ‘natural’ vs. ‘artificial’ fixation), we demonstrate that building features (e.g., transparent interfaces) can attract restorative gaze when optimally designed. An orderly and coherent arrangement of plant and architectural elements provides superior restorative benefits compared to isolated layouts. Thus, designing waterfront spaces to attract attention and guide gaze effectively can improve restoration outcomes.

Research shows that individual visual behavior in urban waterfront spaces is influenced by environmental features. Through intentional design, individuals can be guided to focus on features that positively impact psychological restoration, thereby enhancing restorative benefits.

5.4. Principles for the Restorative Optimization Design of Urban Waterfront Spaces

Based on the research findings, three key design principles are identified: regulating the visibility threshold, optimizing spatial features, and guiding visual behavior. When enhancing environmental features, priority should be given to regulating visual thresholds of water bodies, plants, and architectural elements. In cases where thresholds are restricted, adjustments to the sense of enclosure and interface permeability are necessary to maintain visual coherence along the line of sight in the depth direction. This approach facilitates the creation of a rhythmic spatial sequence, thereby improving spatial exploration and psychological comfort. Regarding the visual attention mechanism, promoting visual comfort and psychological recovery can be achieved by guiding individuals’ visual behavior and balancing visual cognitive load effectively.

5.5. Research Limitations

This study has the following limitations:

(1)

The use of static two-dimensional images and controlled laboratory settings may not fully replicate the complexities of perception and interaction in real-world environments. Future research could enhance the study by incorporating real-world scenarios and utilizing eye-tracking technology to provide additional validation, thus offering a more comprehensive understanding of the spatial restoration mechanism.

(2)

Due to the limitations of the research sample, the generalizability of our findings is limited to young adults (18–30 years) in urban settings. Future studies should examine whether similar design principles apply to other age groups.

(3)

Gender imbalance (with a predominance of females) is a major limitation of the representativeness of the sample in this study. Future studies should employ quota sampling to ensure balanced gender representation, allowing for more definitive conclusions and enhancing the external validity of the findings.

(4)

The focus of this paper is to establish and validate a research framework for restorative studies of urban waterfront spaces, rather than to provide detailed design proposals. Future research can use this as a target to refine the feasibility of these principles.

6. Conclusions

This study focuses on the practical considerations involved in creating restorative urban spatial environments, with a specific focus on the waterfront area of Guangzhou Long Bund. Utilizing multimodal experimental research, this study elucidates the relationship between the environmental features of urban waterfront spaces and their restorative benefits. It highlights the significant impact of visual behavior in this context and puts forth specific spatial optimization design principles.

This study establishes and validates an integrated research framework that systematically links key waterfront design dimensions (elements, morphology, and sequences) with restorative environment theory and combines subjective and objective restorative assessments and eye-tracking methods to elucidate psychological recovery mechanisms. Applying this comprehensive framework, we have accomplished the following:

(1)

Clarified and quantified the complex “environmental features—restorative benefits” influence mechanism within urban waterfront spaces. Specifically, the framework enabled us to accomplish the following: identify critical proportional thresholds of elements (particularly plant and building visibility rate), pinpoint key environmental features impacting benefits across the three design dimensions, and uncover the crucial mediating role of eye movement behavior (e.g., fixation time in the middle-view focal area; balanced distribution of fixation hotspots). The findings demonstrate that element visibility rate and visual characteristics exert the most direct influence, while spatial enclosure and sequence features serve regulatory functions.

(2)

Revealed actionable pathways for design optimization. Based on the mechanisms elucidated by the framework, we propose the optimization principle of “regulating the visibility threshold, optimizing spatial features, and guiding visual behavior,” providing concrete theoretical guidance for restorative design.

This approach offers a quantifiable and actionable foundation for enhancing the restorative design of waterfront areas. Subsequent studies can leverage this methodology to delve deeper into the subject, advancing research on restorative interventions in urban waterfront settings.