Research on Spatial Perception in Virtual Historical Streets Based on Eye-Tracking and Physiological Sensing Data

Abstract

Historical streets are an important component of urban form and embody urban culture. A detailed quantitative analysis of visitors’ spatial perception in historical streets is crucial for enhancing the spatial quality of streets. Focusing on Gulangyu Island, Xiamen, this study proposed a physical–physiological–psychological research framework for spatial perception by constructing a 3D virtual geographic environment. Based on environmental behavior theories and five dimensions of street design quality, it assessed the spatial features of historical streets, visitors’ eye-tracking and physiological sensing data, and psychological perception, summarized the mechanisms of how historical streetscapes influence visitors’ perception, and finally proposed corresponding streetscape optimization strategies. The main findings are as follows. 1) Spatial features of historical streets, including architectural style, layout of commercial spaces, spatial scale, and interface transparency, directly affect visitors’ visual experience and preferences. 2) Visual attention is significantly positively correlated with the historical streets’ imagery, openness, transparency, and complexity, and significantly negatively correlated with enclosure; among these factors, street openness, vitality, and lighting are key factors influencing visitors’ physiological responses. 3) The physical–physiological–psychological interaction mechanisms show that the visual attractiveness and emotional stimulation of the streetscapes can significantly influence individual perception and behavioral decisions, which confirmed the “physical environment–eye-tracking fixation–emotional arousal” mechanism of visitors’ visual preferences. Finally, spatial perception optimization strategies for different types of streets are proposed to inform decision-making for historical street renewals.

Graphical abstract

Keywords

3D Virtual Geographic Environment / Historical Streets / Spatial Perception / Physical–Physiological–Psychological Research Framework for Spatial Perception / Environmental Behavior Theory / Gulangyu Island

Highlight

	· Competing interests  The authors declare that they have no competing interests.
	· Builds a 3D virtual geographic environment for historical street perception with a physical–physio–psych framework
	· Architectural style, layout of commercial spaces, and spatial openness affect visual attention and emotions
	· Summarizes the physical–physio–psych interaction mechanisms of spatial perception in historical streets

1 Introduction

With the advancement of urbanization in China, the focus of urban planning has shifted from spatial growth to quality improvement^[1]. The built-up environment of historical streets not only has an important influence on visitors’ walking behaviors^[2], but also embodies local historical memories and cultural characteristics, playing a significant role in heritage conservation and tourism development. However, historical streets are also facing challenges related to functional changes^[3], which not only alters the physical environment of the streets but also profoundly impacts visitors’ sense of place attachment^[4]. Therefore, studying visitors’ spatial perception and related influencing factors in historical streets is of great theoretical and practical significance for the organic renewal of historical streets and the enhancement of the tourism experience.

2 Literature Review

Historical streets, as significant cultural heritages, have attracted widespread scholarly attention. Scholars have explored the walkability and visual appeal of historical streets using methods such as virtual reality, eye-tracking, and semantic differential. For example, Cheng Shi et al. used virtual reality and decision tree algorithms to precisely measure walkability, providing practical insights for urban micro-renewal^[5]; Zhang Zhang et al. conducted field surveys to reveal the relationship between architectural scale, facade composition, and visitors’ walking and stopping behaviors, offering optimization suggestions for historical street design^[2]; Yuan Li et al. conducted eye-tracking experiments to analyze the complex relationship between visual appeal and visitors’ perceptions of commercial historical streets^[6]. There are also studies focusing on using 3D virtual geographic environments and GIS technology for the display and management of historical streets. For instance, Luís Marques et al. proposed a method for 3D modeling and visualization of cultural heritage based on mobile platforms and augmented reality technology^[7]; Yu-Pin Ma utilized BIM and 3D modeling to improve the cultural heritage preservation process^[8]; Jingxian Tang et al. utilized Tencent street view images, integrating SegNet machine learning with 2D and 3D analysis, to assess the visual quality of hutongs in Beijing, including greenery, openness, and enclosure ^[9]. Overall, both domestic and international research focuses on the physical representation of historical street environments. Methods like virtual reality and eye-tracking experiments explore human-scale spatial perception and behaviors, providing both theoretical and technical bases for the future protection and optimization design of historical streets.

Environmental psychology theory focuses on the relationship between individuals and the environment, particularly the mechanisms through which environmental factors influence individual psychology^{[10] [11]}. In recent years, emerging technologies, such as eye-tracking^[12], wearable cameras^[13], physiological sensors^[14], and virtual reality^[15], have enabled environmental psychologists to more accurately and intuitively analyze individuals’ visual attention^[16], spatial perception ^[17], and real-time emotional responses to street environments^[18], providing strong support for human-scale street renewal design^[19].

In research on spatial perception, scholars have examined the spatial perception measurements of streets from multiple perspectives. At the physical level, Reid Ewing and Susan Handy applied expert rating method to measure the quality of urban street design by proposing an evaluation framework consisting of five dimensions: imageability, enclosure, human scale, transparency, and complexity^{[20] [21]}. This framework has been widely recognized as a critical tool for assessing and enhancing the quality of urban street design. From a physiological perspective, Zheng Chen et al. integrated spatio-temporal trajectory data with affective parameters, effectively mapping participants’ emotional responses to given spaces ^[18]. This approach provides a new perspective for environmental design research. From a psychological standpoint, eye-tracking technology has been extensively applied in studies on visitors’ landscape preferences[22], as well as preferences for environmental elements like street buildings ^[23]~[25].

With advancements in 3D virtual geographic environments^{[26] [27]}, machine learning^{[28] [29]}, and multi-sourced big data^[30], researchers can now more easily construct virtual reality experimental settings^[31] and perform semantic segmentation of eye-tracking attention elements^[6]. These technological developments not only overcome the limitations of previous studies where street perception experiments were often subject to pedestrian-related factors^{[14] [23]}, but also help generate low-cost, controlled 3D virtual geographic environments^[32]. Furthermore, they enable a more detailed representation of historical street features and micro-scale environments^[33], supporting behavior modeling^[34] and integrated data analysis.

In summary, although existing street perception evaluation methods and the application of new data and technologies have provided valuable insights into the evaluations of historical street perception, there is still a lack of studies that combine eye-tracking and physiological sensing within virtual scenarios of historical streets. Based on environmental behavior theories and the five dimensions of street design quality, this study proposes a research framework of spatial perception in historical streets within a virtual geographic environment. It innovatively integrates eye-tracking technology with virtual reality to precisely capture visitors’ visual attention and to explore spatial perception, informing the spatial optimization design of historical streets.

3 Research Methodology

3.1 Research Framework

This study proposed a physical–physiological–psychological research framework for spatial perception on historical streets. At the physical level, spatial attributes and environmental characteristics of historical streets were calculated; five indicators, i.e. imageability, enclosure, human scale, transparency, and complexity, were used to evaluate street spatial features. At the physiological level, eye-tracking and physiological sensing experiments were conducted in a virtual geographic environment; physiological indicators including fixation duration, pupil diameter, galvanic skin response, heart rate, and heart rate variability were collected to analyze the impact of varied spatial factors on perception. At the psychological level, post-experiment evaluations were conducted to analyze participants’ detailed perception on different spatial factors, and then, spatial optimization suggestions for historical streets were proposed (Fig.1).

3.2 Study Area

Gulangyu is an island located in the southwestern corner of Xiamen Island, China, covering an area of approximately 1.88 km². The island’s road system is primarily composed of 35 historic streets. The streets are famous for their historical characteristics, and visitors primarily experience them on foot^[35]. Based on the Historical and Cultural District Protection Plan for Gulangyu Island^[35], preliminary field survey results, and relevant studies^[36][37], a route that links up the key architectural heritages of Gulangyu at maximum was selected for research. Four street types are included on the route: traffic, commercial–residential, residential, and mixed-use (Tab.1, Fig.2). The total length of the route is approximately 3.5 km, and a virtual tour takes 20 ~ 30 minutes at a normal walking speed.

3.3 Participants

A total of 41 students from Xiamen University were recruited both online and offline as participants for the eye-tracking and physiological sensing experiments. Among them, 73.2% were female and 26.8% were male^①. The experiment was conducted at the Xiamen Key Laboratory of Integrated Application of Intelligent Technology for Architectural Heritage Protection from November 22 to 29, 2022. of the participants, 92.7% reported that they had visited Gulangyu Island before and an analysis of variance (ANOVA) showed that prior visits had no significant impact on the evaluation results. When assessing the virtual scenes’ authenticity, participants assigned an average score of 4.09 (out of 5), indicating that the virtual geographic environment in the experiment can effectively replicate the real-world experience in the historical streets of Gulangyu Island (Fig.3, Fig.4). To eliminate potential interference by visitor images on participants’ gaze distribution and psychological perception, ensuring their focus solely on the physical environment, the virtual scenes used in the experiment did not include any depictions of visitors.

① Most eye-tracking experiments recruited around 30 participants, with only a few studies reaching 60 participants (e.g., Ref. [17]). Therefore, the sample size of this study is considered reasonable and adequate.

3.4 Multi-Sourced Data Collection and Integration

The experiment utilized a Gazepoint GP3 device to simultaneously collect participants’ eye-tracking and physiological sensing data.

Step 1: A 3D virtual geographic environment was built using the Unreal Engine 4 (UE4) game engine^[32]. After signing an informed consent and completing eye-tracking calibration, participants conducted a virtual tour along the research route. They filled out a spatial perception questionnaire evaluating their experience in the virtual environment when completing the tour.

Step 2: Following previous studies that use videos and virtual tours as stimulus materials^{[1] [14] [38]}, the collected eye-tracking and physiological sensor data used in this research were segmented into 10-second intervals, recoding changes in spatial coordinates of each segment and plotting eye-tracking heatmaps and physiological indicator variations.

Step 3: The detailed physical environment of the research route was analyzed by segment using SuperMap’s 3D GIS. Additionally, machine learning-based semantic segmentation of streetscape features was performed using the ADE20K dataset, which contains over 27, 000 finely annotated images and is widely recognized in academic studies on scene understanding^{[29] [39]}.

Step 4: Based on the segmented data obtained from Step 2, further statistical processing was conducted on the physiological sensing data. The time window was set to 10 seconds to smooth extreme values, reducing fluctuations and ensuring the accuracy and reliability of the analysis results.

Step 5: All data were integrated into a GIS platform based on a unified spatial coordinate system. The final dataset comprised 8, 375 spatial coordinate points, along with their corresponding eye-tracking and physiological sensing data, spatial indicators for each street segment, and participants’ responses to the perception questionnaire.

3.5 Indicator Selection and Measurement

3.5.1 Physical Indicators

Based on previous studies^{[2][20][32][38]}, the spatial characteristics of the historical streets in the study area, and computational feasibility, this study selected 15 indicators to characterize the spatial elements of historical streets from the five dimensions of street design quality: imageability, enclosure, human scale, transparency, and complexity (Tab.2).

3.5.2 Physiological Indicators

Referring to previous studies^{[18] [22] [38] [40]}, this study selected the following physiological indicators to identify participants’ attention distribution and emotional arousal responses to historical street spatial elements:

1) Fixation points on the objects of interest during gaze duration (FPOGD): The duration time of fixation points on objects of interest, typically measured in milliseconds. This indicator reflects how the human eye allocates attention to objects of interest in the environment.

2) Pupil diameter (LPD for left pupil diameter and RPD for the right): Controlled by the visual nervous system, this is an indicator of visual arousal levels.

3) Galvanic skin response (GSR): The electrical response of the skin to external stimuli. In emotional research, this indicator is commonly used to measure the intensity of emotional arousal.

4) Heart rate (HR) and heart rate variability (HRV): HR refers to the number of heartbeats per unit of time, while HRV indicates the degree of fluctuation between heartbeats over time. Both indicators are widely used to assess emotional arousal levels.

3.5.3 Psychological Indicators

Upon previous research^[41], this study adopted a questionnaire to collect psychological perception data on streetscape experiences. The questionnaire employed the semantic differential method commonly used in psychological studies, providing 11 pairs of adjectives to describe various streetscape characteristics. Participants were required to evaluate each item and assign values ranging from −3 to 3, where 0 represents a neutral perception level, −1 and 1 represent a moderate perception level, −2 and 2 represent a strong perception level, and −3 and 3 represent a very strong perception level (Fig.5).

3.6 Statistical Methods and Models

3.6.1 Spatial Hotspot Analysis

This study employed the Getis-Ord G_i^* method to analyze whether the eye-tracking attention elements and physiological sensing data exhibit significant spatial clustering. The Getis-Ord G_i^* method is commonly used in spatial hotspot analysis^[42], which calculates the values of surrounding elements for each data point (or area) and determines their statistical significance, so as to assess the intensity and location of spatial clustering. The calculation formulas are as follows^[43]:

(1)G_i^{\ast }=\frac{{\sum }_{j=1}^n{w}_{i,j}{x}_j-\overline{X}{\sum }_{j=1}^n{w}_{i,j}}{S\sqrt{\frac{\left[n{\sum }_{j=1}^n{w}_{i,j}^2-{\left({\sum }_{j=1}^n{w}_{i,j}\right)}^2\right]}{n-1}}},

(2)\overline{X}=\frac{{\sum }_{j=1}^n{x}_j}{n},

(3)S=\sqrt{\frac{{\sum }_{j=1}^n{x}_j}{n}},

where x_j represents the attribute value of element j, w_{i, j} denotes the spatial weight between elements i and j, and n is the total number of elements. The result of G_i^* is a z-score, where a significantly positive z-score indicates that higher values (hotspots) are more intensely clustered; conversely, a significantly negative z-score exhibits lower values (coldspots), meaning a higher degree of clustering.

3.6.2 Discrete Choice Model

This study posits that spatiotemporal factors influencing participants’ visual attention preference are highly complex. Therefore, a discrete choice model was employed in this research to analyze the impact of physiological feedback on visual selection behaviors. Based on existing literature^[44] and data availability, the multinomial logit (MNL) model for participants’ visual preference is expressed as follows^[45]:

(4)P(y=i)=\frac{e^{\beta x}}{\sum e^{\beta x}},

where, the dependent variable y represents the participant’s selection for visual element i. Based on eye-tracking heatmap statistics, the output is classified into two mutually exclusive categories (selected or not selected), with corresponding values of 1 or 0. In the multinomial regression analysis, the independent variable x represents the participant’s physiological sensing data, which is standardized; and the parameter β denotes the coefficient of the independent variable.

4 Research Results

4.1 Spatial Hotspot Analysis

The analysis results (Fig.6) indicate that participants had significantly longer fixation durations when touring commercial– residential streets (e.g., Longtou Road, Anhai Road). This suggests that these streets’ distinct commercial styles and high-information-density environmental elements (e.g., storefront signage and entrance spaces) effectively captured participants’ long-duration visual attention. In contrast, for mixed-use streets like Fujian Road, Sanming Road, and Yongchun Road, the rich historical features and diverse spatial elements (e.g., heritage buildings, courtyards, landmarks) led to more frequent shifts in participants’ gaze, resulting in significantly shorter fixation durations.

For residential streets (e.g., Anhai Road) and mixed-use streets (e.g., Fujian Road), a significant increase in PD was closely associated with core impressive spatial elements, such as the blend of historical and modern architecture, which provided rich visual stimuli and fostered deep emotional connections for participants. Conversely, for Bishan Road and Guxin Road, which are traffic streets surrounded by greenery, the dimmer ambient lighting and low-lying shrubs tended to draw participants’ attention, leading to high-value clustering of PD measurements.

On commercial–residential streets such as Zhonghua Road, Shichang Road, and Longtou Road, participants exhibited a significantly higher GSR. These streets not only include a diverse array of street-front shops and high-density building layouts, but also have brighter lighting and greater visual accessibility, which collectively contribute to a richer visual experience. Similarly, in an eye-tracking study on Nanjing Road, Shanghai, Chen Zheng et al. also revealed how outdoor storefront signage shapes “unique, ” “rich, ” and “distinctive” streetscapes that provided strong visual impression for pedestrians^[12].

High-value HR and HRV clusters were concentrated around Lujiao Road and Fujian Road. The key heritage buildings in these areas not only captured attention from most participants but also reinforced emotional connections by evoking personal memories and experiences. For example, some participants verbally reported that while touring Fujian Road, they recalled their previous visits to the Haitian-Tanggou Mansion.

4.2 Spatial Hotspot Analysis of Visual Attractors

Based on relevant literature^[46] and the spatial features of historical streets in Gulangyu Island, this study categorized visual attractors in participants’ eye-tracking preferences into architectural elements, road elements, natural elements, and landmark elements, comprising a total of 16 subcategories. Among them, architectural elements were further classified into modern building, traditional building, Western-style building, and architectural heritage component, based on their construction period and architectural style^[47] (Tab.3). Road elements include street, square, staircase/slope, and wall/railings; natural elements consist of sky, vegetation, mountain/rocks, and sea; and landmark elements encompass storefront signage, fountain, streetlight, and road sign.

The spatial hotspot analysis of eye-tracking data (Fig.7) shows that high-value clusters of architectural elements were primarily concentrated in zones with a higher density of historical buildings, where distinctive architectural styles exerted a strong visual attraction on participants. Road elements exhibited high-value clusters at street intersections and the locations with clearly defined spatial boundaries, where the high visibility and spatial openness allowed participants to pass through or linger, underscoring their importance in pedestrian flow. Natural elements showed high-value clusters in beachfront and park areas, demonstrating the strong visual appeal of natural landscapes to participants. Meanwhile, landmark and iconic elements formed high-value clusters in commercial–residential streets and spatial nodes like the intersection of Guxin Road and Sanming Road, known as the “most beautiful corner on the island.”

4.3 Correlation Analysis Between Street Spatial Features and Visual Attractors

This study examines the correlation between street spatial features and visual attractors, further revealing the influence of the physical environment on psychological perception (Fig.7, Fig.8).

1) Imageability: Pedestrian activity level showed a negative correlation with street and vegetation but a positive correlation with storefront signage. Meanwhile, the proportion of historic building was positively correlated with Western-style building, architectural heritage component, mountain/rocks, and fountain. These results suggest that in areas with higher pedestrian dynamics, commercial elements such as storefront signage are more likely to attract participants’ visual attention. This is because commercial areas typically feature more interactive spaces and stronger visual stimuli, where visitors are often attracted by advertisements and signage with bright colors and distinctive fashions. This trend was particularly evident in commercial–residential streets such as Longtou Road and Shichang Road. In contrast, streets with a higher proportion of historic buildings, such as Fujian Road and Huangyan Road, became popular focal points for participants due to their rich historical and cultural significance and distinct architectural styles. Western-style buildings and architectural heritage components often have strong historical symbolism, which is particularly prominent in Xiamen—as a city with a unique cultural and historical context, Xiamen’s historical sites and buildings exhibit a greater visual appeal, and the fusion of traditional and Western architectural styles provides a more diverse visual experience for visitors.

2) Enclosure: Street width showed a positive correlation with Western-style building, architectural heritage component, and fountain, while the height-to-width ratio (same side) was positively correlated with square and storefront signage. The street layout and architectural characteristics significantly influenced participants’ visual attention—wider streets would lead to a greater attention to buildings and public spaces, whereas narrower streets would enhance the visual prominence of road elements. For streets with rich heritage buildings, such as Fujian Road, wider streets allowed participants to gaze not only culturally significant buildings but also artificial structures such as fountains. In contrast, for traffic streets such as Sanming Road, narrower streets constrained participants’ sightline, making street itself and vegetation the primary visual attractors. Additionally, a positive correlation was observed between building height and height-to-width ratio with storefront signage, particularly on commercially active streets such as Longtou Road. On these streets, the spatial disparity makes taller buildings and storefront signage more likely to attract visual attention.

3) Human scale: Sky openness was positively correlated with participants’ visual attention to sky, while the volume of the field of view showed a positive correlation with traditional building, mountain/rocks, and fountain. For streets with higher sky openness, such as Fujian Road and Huangyan Road, traditional buildings and natural elements became visual attractors. These elements stood out due to their cultural and natural significance, forming a strong contrast with the surroundings, leading to frequent visual attention from participants. For example, in areas where Sunlight Rock is clearly visible, the unique natural landscape attracted significant visual attention. This finding confirms Gulangyu Island’s identity as an “Architectural Museum of the World, ” which is represented not only in its diverse architectural styles but also in the harmonious integration of natural and built-up landscape elements, such as mountain/rocks and fountain. These elements may have been deliberately considered during the site selection of buildings, or carefully designed in different historical periods. Additionally, for streets with expansive field of view (e.g., Anhai Road), the sky that occupied a larger proportion but contained less detailed information tended to attract more visual attention.

4) Transparency: The proportion of wall was positively correlated with the visual attention to wall/railings. For narrower traffic streets, wall/railings, despite restricting participants’ sightline, became prominent attractors due to their ornamentation and architectural details. The concept of transparency concerns how visual interfaces allow sightlines to penetrate, facilitating visual connection and interaction. In environments with limited visibility, where transparency is lacking, participants’ attention is more likely to be drawn to decorative and detailed interfaces. The importance of transparency for spatial creation is also demonstrated by numerous international urban design practices. For example, in Alexandria, Egypt, specific ground-floor window ratio requirements have been implemented for commercial retail buildings, not only improving the visibility of activities on the street but also enhancing the vibrancy of the streetscape^[48].

5) Complexity: The proportion of building was positively correlated with modern building but negatively correlated with street, vegetation, and sea. Similarly, color contrast gradient showed a positive correlation with modern building and a negative correlation with road sign. For streets where buildings occupied a higher proportion (such as residential streets), modern buildings became main visual attractors due to their vivid color contrasts. In contrast, for streets where road signs were the prominent visual attractors, muted or simple color schemes are often intentionally used to enhance the contrast gradient of these navigational elements, ensuring effective information conveying.

4.4 MNL Model Analysis of Visual Attractors and Physiological Indicators

To explore the relationship between psychological perception and physiological indicators, this study employed an MNL model to quantify the impact of different visual attractors on physiological responses. The results revealed that only 7 visual attractors— modern building, traditional building, street, wall/railings, sky, vegetation, and storefront signage—exhibited significant correlations with physiological indicators (Tab.4).

Physiological responses were not particularly strong when participants fixated on modern buildings. However, fixation duration and HRV exhibited a significant negative correlation. Modern buildings are typically characterized by simplistic, structured, and smooth-lined facades, which lack organic elements and variation, compared with natural landscapes. This visual monotony may lead to reduced emotional arousal, consequently lowering HRV.

When participants fixated on traditional buildings, walls or railings, significant PD changes were observed in both eyes. Previous studies have found that pupil size is generally larger when viewing objects within the area of interest, compared with those outside the area ^[17]. Larger PD might relate to greater visual interest in traditional buildings, walls or railings, while smaller PD could be associated with recalled memories and cognitive load^[49]. That is, visitors may need to consume more attention to process and retrieve memories related to traditional architecture.

Street exhibited a significant positive correlation with GSR and HR. This may be because visual stimuli on the street (e.g., pavement patterns, directional signs) captured participants’ attention, and led to increased GSR and HR.

Wall/railings showed a significant positive correlation with HRV, meaning that higher and denser wall/railings were associated with greater HRV values. This can be explained by the frame effect created by tall and dense barriers, where lower street transparency induced feelings of oppression and discomfort, thereby increasing HRV.

Sky and vegetation exhibited a significant negative correlation with fixation duration. Previous research has confirmed that although sky and greenery often occupy a large portion in the field of view, they contain relatively low information density, leading to shorter fixation durations^{[50] [51]}. Additionally, sky, as an open and calming element, had a significant negative impact on GSR and HR, meaning that observing the sky led to decreases in both physiological indicators.

When participants fixated on storefront signage, smaller RPD, a slight increase in GSR, and a significant increase in HRV were observed. This can be explained by the increased cognitive load required participants to examine the signage in-detail, which induced smaller PD^[49]; while storefront signage stimulated participant interest, leading to increases in both GSR and HRV.

4.5 Regression Analysis of Psychological Perception and Physiological Indicators

Through a preliminary regression analysis of psychological perception and physiological indicators with OLS models, the study found that only fixation duration, PD, and HR exhibited statistically significant correlations with the psychological perception items (enclosed–open, lively–quiet, and dark–bright) (Fig.9). The R² values for these models were 0.14, 0.12, and 0.11, respectively, indicating that Model 1 explained 14% of the variance in fixation duration, Model 2 explained 12% of the variance in RPD, and Model 3 explained 11% of the variance in HR.

In terms of street openness, Model 1 indicated a positive correlation between fixation duration and perceived openness, suggesting that spacious streetscapes led to longer visual fixations. This implies that wide, open spaces may enhance attention and induce pleasant feelings, aligning with the finding by Shanzhi Kang et al., which demonstrated that open spaces contribute to positive emotional experiences^[14].

Regarding street vitality, Model 2 showed that stronger perception of bustling streetscape (indicating the presence of more areas of visual interests on the street) corresponded to larger PD. Previous studies have also confirmed that outdoor signage in the streets enhances visitors’ attention^[50], which in turn leads to larger PD.

In terms of lighting, Model 3 revealed that greater perception of brightness was associated with lower HR, suggesting that well-lit streetscapes promote positive emotions, making the participants more relaxed and calm.

5 Discussion

5.1 Analysis of the Physical–Physiological–Psychological

Examining participants’ spatial perception in historical streets and its influencing factors and based on the above experiment results, this study further summarized the physical–physiological–psychological interaction mechanisms underlying as follows.

1) Physical factors, particularly architectural style, layout of commercial spaces, spatial scale, and interface transparency, directly influence visitors’ visual experience and preferences.

2) Physiological indicators, such as fixation duration, HR, and PD variations, reflect visitors’ real-time emotional responses to different visual attractors in the streets, revealing the potential of local distinctiveness of historical streets in enhancing tourism experiences.

3) Psychological factors serve as a bridge between the physical environment and physiological responses. The correlation analysis between psychological perception and physiological indicators confirmed the “physical environment–eye-tracking fixation– emotional arousal” mechanism of visitors’ visual preferences, highlighting that street openness, vitality, and lighting are key factors in enhancing pedestrian experience.

5.2 Optimization Strategies for Historical Streets

1) Traffic streets. These streets are characterized by numerous walls and facades, abundant greenery, but a visually monotonous spatial experience. To improve visual appeal, it is necessary to increase street width and openness while enhancing wall transparency. For instance, installing interactive facilities and display installations while maintaining traffic efficiency can enhance street vitality.

2) Residential streets. These streets are often narrow, with low openness, dense modern buildings, and limited natural elements. Considering the needs of both local residents and visitors, facade renovation can be implemented to improve street cleanliness and openness. This can be achieved through strategies such as creating small-size green spaces (e.g., pocket parks), integrating the residential streets with the natural environment.

3) Commercial–residential streets. As places attracting visitors’ attention, these streets feature varied storefront signage and lighting interplay. Optimization measures include enhancing lighting control to prevent physiological discomfort caused by dramatic lighting fluctuations of signage; and improving business management for a better integration of modern commerce and historical buildings.

4) Mixed-use streets. Such streets densely accommodate cultural architectural heritages. However, due to heritage protection requirements, they often have massive walls and facades with low transparency and weak interaction. To enhance visual appeal while preserving historical structures, measures of increasing the transparency of street interfaces can be taken. For example, by leveraging VR and AR technologies to reconstruct historical scenes and buildings, or to introduce digital heritage exhibitions and NPC-guided interactions, visitors’ attention and engagement would be enhanced.

6 Conclusions

This study employed an eye-tracking and physiological sensing experiment to explore participants’ spatial perception in historical streets and related influencing factors, capture individual visual attention preferences, and uncovered the physical–physiological– psychological interaction mechanisms underlying participants’ spatial perception in historical streets. The findings indicate that, at the physical level, architectural and commercial elements in historical streets significantly influence visitors’ spatial perception. Specifically, streetscapes characterized by high imageability, low enclosure, high openness, high transparency, and high complexity attract greater visual attention and elicit stronger emotional responses. At the physiological level, HR and PD reveal visitors’ visual preferences for different street types. In traffic streets, walls and railings were the dominant visual attractors; in commercial– residential streets, storefront signage was the primary visual attractor; in residential streets, modern buildings attracted more attention; and in mixed-use streets, participants exhibited a preference for traditional buildings, particularly architectural heritage components and natural elements. At the psychological level, the perceived street openness, vitality, and lighting influenced participants’ physiological responses, demonstrating that the visual appeal and emotional arousal of the environment significantly affect individuals’ visual preferences.

The physical–physiological–psychological research framework for spatial perception proposed in this study enriches the application of environmental psychology theories in historical street research and offers a novel perspective for studies on pedestrian-friendly historical streets. Methodologically, this study developed a 3D virtual geographic environment and an integrated eye-tracking and physiological sensing experiment, providing a new quantitative approach to exploring environmental perception in historical streets and improving the precision of research results. In terms of application, the study offers human-centered urban renewal strategies for the four types of historical streets in Gulangyu Island.

Despite its contributions, this study has some limitations. Future research should further assess the fidelity of the 3D virtual geographic environment and enhance experimental fluidity, such as adopting head-mounted VR eye-tracking systems. More historical streets should be studied with the research framework proposed in this paper, and deep learning techniques can be utilized to enhance the accuracy and efficiency of data analysis, refining optimization strategies for historical streets.

References

[1] Wu, Y. , Dong, W. , & Zhang, W. (2022) Research on influencing factors of built environment perception in neighborhoods: Evidence from behavioral experiment. City Planning Review, 46 ( 12), 99– 109.

[2] Zhang, Z. , Xu, G. , Li, W. , Long, Y. , & Cao, Z. (2019) The impact of the micro-scale built environment of historic street on visitor’s walking behaviors: A case study on Wudaoying Hutong in Beijing. Architectural Journal, ( 3), 96– 102.

[3] Wu, Z. , Zhang, L. , Zheng, Z. , Zhang, B. , & Liao, Z. (2019) The production of space within the field of tourism: A case study of the tourism community of Chaozhou ancient city. Tourism Tribune, 34 ( 12), 86– 97.

[4] Liu, J. , Zhang, M. , Zhang, X. , & Hong, X. (2023) Relationship between soundscape and place attachment of tourists in historic block: A case study of Jinli Ancient Street. Chinese Landscape Architecture, 39 ( 1), 59– 64.

[5] Shi, C. , Yuan, Q. , Pan, H. , & Ye, Y. (2020) Measuring walkability of street space and its design control in the context of new analytical techniques: A case study of Shanghai Jing’an Temple area. Shanghai Urban Planning Review, ( 5), 71– 79.

[6] Li, Y. , Huang, J. , Liang, J. , Zhang, Y. , & Chen, Y. (2022) Research on visual attraction and influencing factors of perception of commercial street space in cultural heritage site: Taking Gulangyu Longtou Road as an example. Journal of Human Settlements in West China, 37 ( 2), 114– 121.

[7] Marques, L. F. , Tenedório, J. A. , Burns, M. , Romão, T. , Birra, F. , Marques, J. , & Pires, A. (2017) Cultural heritage 3D modelling and visualisation within an augmented reality environment, based on geographic information technologies and mobile platforms. ACE: Architecture, City and Environment, 11 ( 33), 117– 136.

[8] & Ma, Y. (2021) Extending 3D-GIS district models and BIM-based building models into computer gaming environment for better workflow of cultural heritage conservation. Applied Sciences, 11 ( 5), 2101– .

[9] Tang, J. , & Long, Y. (2019) Measuring visual quality of street space and its temporal variation: Methodology and its application in the Hutong area in Beijing. Landscape and Urban Planning, ( 191), 103436– .

[10] & Li, B. (2008) Environmental behavior theory in environmental behavior studies and its expansion. Architectural Journal, ( 2), 30– 33.

[11] Zhuang, C. , & Zhang, J. (2011) Place identity: Analysis from environmental psychology perspective. Advances in Psychological Science, 19 ( 9), 1387– 1396.

[12] Chen, Y. , Chen, Z. , & Du, M. (2022) Designing attention—Research on landscape experience through eye tracking in Nanjing Road Pedestrian Mall (Street) in Shanghai. Landscape Architecture Frontiers, 10 ( 2), 52– 70.

[13] Zhang, Z. , & Long, Y. (2019) Application of wearable cameras in studying individual behaviors in built environments. Landscape Architecture Frontiers, 7 ( 2), 22– 37.

[14] Kang, S. , Chen, X. , Huang, C. , Li, J. , Yan, J. , & Ye, Y. (2023) The exploration of street micro-renewal technology from the perspective of new urban science: A case study of Guzong Road in Nanhui New City. New Architecture, ( 6), 52– 57.

[15] Chen, B. , Gong, C. , & Li, S. (2022) Looking at buildings or trees? Association of human nature relatedness with eye movements in outdoor space. Journal of Environmental Psychology, ( 80), 101756– .

[16] Zhang, R. , Li, G. , Zhang, L. , & Huang, Q. (2023) Research on visual perception of historical district’s cityscape by VR eye-tracking technology: A case study of Zhangzhou Ancient City. New Architecture, ( 6), 58– 63.

[17] Liu, L. , Wu, M. , Ma, Y. , & Qu, H. (2021) A review of eye tracking researches in landscape architecture field. Journal of Human Settlements in West China, 36 ( 4), 125– 133.

[18] Chen, Z. , & Liu, S. (2018) Real-time environmental affective experience assessment via wearable sensors. Chinese Landscape Architecture, 34 ( 3), 12– 17.

[19] Ye, Y. , & Dai, X. (2017) Spatial perception and design potentials in the context of new analytical techniques and new data. Time+Architecture, ( 5), 6– 13.

[20] Ewing, R. , Handy, S. , & Jiang, W. (2012) Measuring the unmeasurable: Urban design qualities related to walkability. Urban Planning International, 27 ( 5), 43– 53.

[21] Ewing, R. , & Handy, S. (2009) Measuring the unmeasurable: Urban design qualities related to walkability. Journal of Urban Design, 14 ( 1), 65– 84.

[22] & Ren, X. (2019) Consensus in factors affecting landscape preference: A case study based on a cross-cultural comparison. Journal of Environmental Management, ( 252), 109622– .

[23] Hollander, J. B. , Sussman, A. , Purdy Levering, A. , & Foster-Karim, C. (2020) Using eye-tracking to understand human responses to traditional neighborhood designs. Planning Practice & Research, 35 ( 5), 485– 509.

[24] Noland, R. B. , Weiner, M. D. , Gao, D. , Cook, M. P. , & Nelessen, A. (2017) Eye-tracking technology, visual preference surveys, and urban design: preliminary evidence of an effective methodology. Journal of Urbanism: International Research on Placemaking and Urban Sustainability, 10 ( 1), 98– 110.

[25] Shemesh, A. , Leisman, G. , Bar, M. , & Grobman, Y. J. (2022) The emotional influence of different geometries in virtual spaces: A neurocognitive examination. Journal of Environmental Psychology, ( 81), 101802– .

[26] Lin, H. , Hu, M. , Chen, M. , Zhang, F. , You, L. , & Chen, Y. (2020) Cognitive transformation from geographic information system to virtual geographic environments. Journal of Geo-Information Science, 22 ( 4), 662– 672.

[27] Liu, H. , & Xue, M. (2021) Exploring the use of virtual geographic environments for geo-spatial cognition research. National Remote Sensing Bulletin, 25 ( 10), 2027– 2039.

[28] Shao, Y. , Ye, D. , & Ye, Y. (2023) Measuring the transparency of street interface based on street view images and deep learning: Taking Shanghai as an example. Urban Planning International, 38 ( 6), 39– 47.

[29] Li, X. , Shang, W. , Liu, Q. , Liu, X. , Lyu, Z. , & Ochieng, W. (2024) Towards a sustainable city: Deciphering the determinants of restorative park and spatial patterns. Sustainable Cities and Society, ( 104), 105292– .

[30] & Ye, Y. (2019) The new potentials of urban design in the context of new urban science. Journal of Human Settlements in West China, 34 ( 1), 13– 21.

[31] Xu, L. , Meng, R. , Huang, S. , & Chen, Z. (2019) Healing oriented street design: Experimental explorations via virtual reality. Urban Planning International, 34 ( 1), 38– 45.

[32] Li, Y. , & Guo, J. (2023) Research on heritage value perception of historical streetscapes based on 3D virtual geographic environment under the background of digital cultural tourism: Taking Gulangyu Island in Xiamen as an example. China Ancient City, 37 ( 6), 50– 59.

[33] Wu, Z. , & Ai, X. (2020) Research of spatial visual features of Quanzhou Dacuo houses based on 3D GIS. Journal of Huaqiao University (Philosophy & Social Sciences), ( 2), 140– 150.

[34] Wu, C. , Zhou, S. , & Ye, Y. (2021) The application of human behaviour modelling in urban ergonomics design. World Architecture, ( 3), 46– 49.

[35] Xiamen Municipal People’s Government. (2020). Gulangyu Historic and Cultural District Conservation Plan.

[36] Li, Y. , Guo, J. , & Chen, Y. (2022) A new approach for tourists’ visual behavior patterns and perception evaluation based on multi-source data. Journal of Geo-Information Science, 24 ( 10), 2004– 2020.

[37] Huang, J. , Liang, J. , Yang, M. , & Li, Y. (2024) Street space visual quality evaluating method of tourism sites based on street view images. Journal of Geo-Information Science, 26 ( 2), 352– 366.

[38] Pei, W., Guo, X., & Lo, T. (2021). Biofeedback in the dynamic VR environments: A method to evaluate the influence of architectural elements on human spatial perception. WABD 2021: 2021 Workshop on Algorithm and Big Data.

[39] Yao, Y. , Liang, Z. , Yuan, Z. , Liu, P. , Bie, Y. , Zhang, J. , Wang, R. , Wang, J. , & Guan, Q. (2019) A human-machine adversarial scoring framework for urban perception assessment using street-view images. International Journal of Geographical Information Science, 33 ( 12), 2363– 2384.

[40] Ge, Y. , Chen, Y. , Liu, Y. , Li, W. , & Sun, X. (2014) Electrophysiological measures applied in user experience studies. Advances in Psychological Science, 22 ( 6), 959– 967.

[41] Li, Y. , Yang, L. , & Gao, X. (2019) Preference and spatial improvement strategies of street space of Gulangyu Island. Planners, 35 ( 14), 24– 31.

[42] García-Palomares, J. C. , Gutiérrez, J. , & Mínguez, C. (2015) Identification of tourist hot spots based on social networks: A comparative analysis of European metropolises using photo-sharing services and GIS. Applied Geography, ( 63), 408– 417.

[43] Ord, J. K. , & Getis, A. (1995) Local spatial autocorrelation statistics: Distributional issues and an application. Geographical Analysis, 27 ( 4), 286– 306.

[44] Ye, Y. , Zhou, X. , & Wang, Z. (2019) A quantitative approach for measuring positive social impacts of tall buildings’ lower public space: Utilizing virtual reality and biosensors techniques. Time+Architecture, ( 6), 152– 159.

[45] Wang, C. , Wang, D. , Zhu, W. , & Song, S. (2015) Research progress of discrete choice models. Progress in Geography, 34 ( 10), 1275– 1287.

[46] Zhang, R. , & Yin, B. (2020) Research on the influence of environmental perception of guiding path preference under spatial value matching: A case study of Gulangyu. World Architecture, ( 11), 115– 119.

[47] Wang, J. , & Chen, Z. (2020) Study on the change of gate building style in modern Gulangyu. Architecture & Culture, ( 6), 134– 137.

[48] Li, X. , Li, Y. , Ren, Y. , & Lask, A. (2020) Research on the urban spatial visual quality based on subjective evaluation method and eye tracking analysis. Architectural Journal, ( S2), 190– 196.

[49] Kucewicz, M. T. , Dolezal, J. , Kremen, V. , Berry, B. M. , Miller, L. R. , Magee, A. L. , Fabian, V. , & Worrell, G. A. (2018) Pupil size reflects successful encoding and recall of memory in humans. Scientific Reports, ( 8), 1– 7.

[50] & Chen, Z. (2023) Positive nudges for urban regeneration via eye-tracking and behavioural evidence. World Architecture, ( 7), 68– 69.

[51] Jin, Y. , Kuang, X. , Xi, T. , & Chen, Z. (2023) Accurate transformation of street greening promoted by decision-making on health behavior. Landscape Architecture, 30 ( 1), 45– 53.

RIGHTS & PERMISSIONS

© Higher Education Press 2025