Health tourism model in the digital age: emotional healing effects of disembodied landscape perception through social media

Abstract

In the digital era, increasing attention has been paid to how social media reshapes virtual tourism, yet little research has explored the mechanisms of digital healing in disembodied travel and the differentiated effects of landscape types. This study takes Erhai Scenic Area in China as an example and combines text coding with questionnaire survey. Study 1 used web crawling to collect review texts, followed by semantic analysis and landscape coding. Study 2, based on the classifications derived from Study 1, conducted video experiments and questionnaire surveys. The results of both studies were then compared and analyzed. Findings indicate that emotion plays a partial mediating role in the mechanism of "disembodied landscape perception—embodied emotional change—embodied travel intention," with personality traits (openness and neuroticism) serving as moderators. Moreover, the four landscape types: ecological-natural, commercial-leisure, historical-cultural, and rural-pastoral demonstrate differentiated effects in emotional healing, promoting travel intentions, and facilitating the transformation from disembodied to embodied perception. This study provides a theoretical foundation for destination marketing, healthy landscape design, and digital healing, contributing to the development of new health tourism models in the social media context.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36305-8.

Introduction

In the context of an era where virtual technologies are deeply embedded in daily life, human spatial experiences exhibit a dual characteristic of both “embodied” and “disembodied” elements. Currently, social media, short video platforms, and virtual technologies are creating a new tourism paradigm where “disembodied” experiences on social media (such as travel photos, videos, movies, etc.) inspire “embodied” travel intentions. This “disembodied tourism” acts as a precursor or substitute for physical travel, fulfilling a growing desire for escapism amidst the constraints of modern, high-pressure lifestyles, providing a new type of travel consumption model. Embodied Cognition Theory (ECT) posits that cognition is shaped by bodily states, memory, and the interaction between the body and the environment¹. It is highly dependent on the characteristics of the body², providing a theoretical foundation for analyzing tourist experiences from the perspective of “bodily presence,” as well as how these experiences shape destination image perception³, emotional responses, and behavioral intentions⁴. Most existing studies⁵ focus on the immediacy of offline embodied experiences⁶, where the interaction between body and environment influences emotions and behaviors¹, yet they tend to overlook how digital media can, through “quasi-embodiment” mechanisms, evoke similar emotional responses⁷. This study challenges the traditional notion that embodied experiences in tourism must rely on the absolute physical presence of the body, extending ECT into the digital and virtual domain, and thereby laying the groundwork for future research on perceptual experiences in disembodied contexts. With the advancement of digital technologies, social media has introduced image-based virtual interactions that replace physical contact, shifting social activities toward disembodied experiences⁸. Disembodied experience refers to a state in which mental cognition is somewhat separated from the physical body⁹. In digital contexts, this is typically achieved through interfaces⁸, avatars, or audiovisual content¹⁰, which can effectively evoke and regulate emotions, triggering genuine affective reactions. AI companions can simulate empathy through affective algorithms, creating interactive experiences with emotional warmth¹¹. Experiments involving two-dimensional virtual hands and VR have demonstrated that virtual body representations can be flexibly adjusted to a disembodied state¹⁰. Research on virtual identity and consumption has enabled anonymous, bodiless interactions^12–14. Media such as painting, film, video, and games immerse viewers in simulated worlds¹⁵. For example, the game “Black Myth: Wukong” provides a two-dimensional, disembodied experience that stimulates the senses¹⁶. Such virtual interactions may also induce changes in bodily perception⁷, while enhancing environmental awareness and engagement¹⁷. These technologies eliminate physical barriers to evoke a “pseudo-presence,”^17,18 indicating that physical presence is not a prerequisite for shaping emotions. Digital technologies can activate body schemata and emotional memory by simulating embodied mechanisms, thereby producing effects comparable to those of actual embodied experiences.

Health tourism is a dynamic phenomenon¹⁹,which is generally understood as tourism in which individuals travel to other places to engage in health-related activities, including both disease prevention and medical treatment²⁰. The model of health tourism examined in this study, however, is defined by its core function of promoting physical and mental well-being. Within the context of digital media, the scope is further extended to encompass virtualized disembodied experiences. Landscape perception, as an important theoretical foundation for the development of health tourism, refers to tourists’ subjective interpretation of and emotional response to the environment. It constitutes the core of the tourism process, directly influencing satisfaction and health outcomes²¹. The concept of “restorative environmental perception” explains how specific environments promote emotional recovery and stress relief²². A large body of research indicates that exposure to natural environments, such as forests and waterscapes, can reduce environmental stressors²³ and alleviate negative emotions²⁴. Coastal environments, in particular, have been shown to exert positive effects on both physical and mental well-being²⁵, demonstrating the health and wellness benefits of nature-based tourism for humans²⁶. In the post-pandemic era, tourists’ attention to health has significantly increased, with a tendency to choose health tourism as a means to improve physical and mental well-being²⁷. Within the process of health tourism, the core objective is the therapeutic effect on both body and mind. Contemporary scholars define “healing” as the holistic restoration of individuals from a state of imbalance, encompassing both physiological recovery and subjective psychological perception^28,29. As an emerging paradigm, digital healing refers to the mechanisms through which digital media content or interactive experiences positively influence users’ emotional states, alleviate stress, and facilitate psychological recovery. Research has demonstrated that disembodied experiences possess psychological restorative effects, enhance psychological resilience, and improve environmental perception. Creative video therapy, building on cinema therapy and neurocinematics, unlocks the healing potential of media³⁰. Animated videos can soothe and heal psychological wounds³¹, while pet and food-related content facilitates emotional release and reduces distress³². Studies confirm that visual and tactile stimuli positively impact emotions³³. Digital art environments and virtual reality also elicit therapeutic emotional responses³⁴. During COVID-19, virtual social interaction was shown to enhance psychological resilience³⁵ and reduce virus-related anxiety³⁶. These studies highlight the benefits of digital healing in promoting mental health and create new opportunities for exploring the healing functions of landscape perception within health tourism.

Current research on the therapeutic effects of landscape perception on psychological well-being is already extensive; however, much of the literature primarily focuses on on-site experiences, while relatively little attention has been paid to the digital or disembodied travel contexts. Bastiaansen et al. employed electroencephalography to measure tourists’ emotions when viewing destination videos and images³⁷, whereas Zhang et al. utilized computer-based systems to detect emotional differences among tourists³⁸. Since emotions can be triggered by external stimuli³⁹, the online environments created by virtual tourism can effectively alleviate stress⁴⁰. Research on virtual cultural heritage tourism has explored the positive impact of VR sensory technologies on users’ emotional states⁴¹. Experiments with digital museums demonstrate that digital technologies preserve cultural authenticity while enhancing therapeutic potential⁴². Experimental evidence from VR further confirms that video experiences can facilitate emotional healing⁴³, providing empirical support for the industrial development of therapeutic tourism⁴⁴. In addition, personality traits can influence emotion regulation strategies and cognitive processing⁴⁵. The Big Five personality traits have been shown to be associated with various emotion regulation mechanisms, including reappraisal and mindfulness⁴⁶. Specifically, openness has been demonstrated to moderate the relationship between travel videos and travel inspiration⁴⁷, whereas neuroticism affects emotional responses under different arousal conditions⁴⁸. Research on the therapeutic effects of health tourism has been steadily deepening, confirming the close connection between travel videos and viewers’ emotions. Nevertheless, most existing studies have approached this from the perspective of a single landscape or overall environment, with few studies deconstructing the video content itself to explore how the landscapes and activities presented influence viewers’ emotions⁴⁹. This has led to a lack of differentiation among landscape types, as well as the absence of quantitative comparisons of the therapeutic effects of different landscape types in both real and video environments. Based on this, this study approaches the topic from the perspective of landscape perception in travel videos and propose the following H1. In Study 1, H1 was further refined based on landscape analysis results, comparing the differential effects of perceived landscape types on emotional changes. At the same time, considering the moderating role of personality traits (openness and neuroticism) in this process, we propose the following hypotheses:

H1

The landscape perception brought by disembodied travel video has a positive impact on people’s emotional changes.

H2a

Openness positively moderates the impact of landscape perception to emotional changes.

H2b

Neuroticism positively moderates the impact of landscape perception to emotional changes.

According to the cognitive-behavioral approach⁵⁰, travel intention is a key factor in shaping travel experiences. Although virtual tourism can partially substitute for on-site travel, expectations of actual travel experiences typically exceed the visual imagination of the objects⁵¹. Embodied travel intention refers to the degree of effort users are willing to invest in real-world travel activities corresponding to their virtual tourism experiences⁵². Research has shown that tourists’ attitudes toward smartphones influence travel intention, which in turn affects immersive engagement⁵³. Positive sensory experiences related to destinations on social media increase tourists’ attachment to and identification with those destinations⁵⁴. In today’s digital environment, short-form travel content is widely prevalent, and its design, along with destination attractiveness, can enhance viewers’ desire to travel⁴⁷. Sensory-rich culinary tourism videos can elicit emotions and stimulate behavioral intentions¹⁶. Moreover, viewers’ affinity for travel videos can generate emotional connections that motivate them to visit the destination⁵⁴. Virtual tourism experiences may have a positive influence on subsequent behavioral decisions, as travel videos can simulate real-life scenarios to evoke travel experiences⁵⁵. Recent studies have begun to explore the transition between embodied and disembodied tourism, aiming to clarify the mechanisms linking disembodied landscape perception, emotional changes, and travel intention, and to examine whether emotional changes play a mediating role, we propose the following three hypotheses and establish the basic framework of this study (Fig. 1):

Fig. 1.

Basic mechanism framework (self-drawn by the author).

H3

Disembodied landscape perception has a positive influence on embodied travel intention.

H4

Emotional changes positively influence travel intention.

H5

Emotional changes have a mediating effect on the relationship between disembodied landscape perception and embodied travel intention.

Study 1

Research area

Erhai Lake Scenic Area, located in Dali City, Yunnan Province, boasts rich natural and cultural resources (Fig. 2). As China’s seventh largest freshwater lake, it serves as a key ecological zone. According to the results of a survey on Chinese tourists’ ideal health tourism destinations, the top three health tourism destinations are Hainan Province, Yunnan Province, and Guangxi Province²⁷. With rural tourism driving rural revitalization, tourism around Erhai Lake has rapidly expanded⁵⁶. Development began in 1997 with the Dali Bai Autonomous Prefecture Tourism Industry Development Group, attracting general tourists after 2007 and peaking in 2016⁵⁷. In recent years, tourism has surged again. The area’s pastoral scenery and traditional settlements, combined with Erhai’s natural environment⁵⁸, form a strong foundation for Dali’s tourism appeal and have made it a focus of domestic tourism research. Choosing the Erhai Lake Scenic Area as the research area provides a rich sample of landscape perception types, making it highly representative.

Fig. 2.

Location analysis map (self-drawn by the author, base map satellite image from Amap, https://ditu.amap.com/).

Data source

The data for this study was collected from online reviews on tourism websites, focusing on the perceptions of visitors after experiencing the sites. To ensure the comprehensiveness and representativeness of the research data, genuine perception evaluations from visitors were analyzed. The data was gathered from multiple channels, including three popular Chinese travel websites: Ctrip, Qunar, and Mafengwo.

Methods of study 1

Scholars have employed various landscape classification methods based on principles such as functional morphology, dominant factors, and scale attributes to partition large-scale spaces. Focusing on tourists’ perceptions of the Erhai Scenic Area, this study notes that conventional classification approaches lack regional specificity and perceptual relevance. To better refine tourists’ embodied landscape perception types and translate these into disembodied perception, this study innovatively re-codes landscape classification by using online review data. Through NVivo qualitative analysis, a three-level classification of spatial landscape perception characteristics for the Erhai Scenic Area was derived. Study 1 integrated online text semantic analysis and Grounded Theory coding methods^59,60 to extract the main types of tourists’ landscape perceptions of the Erhai Scenic Area from evaluation texts after embodied tourism, and conducted sentiment analysis on text evaluations of different types of landscapes. This provided a classification basis for exploring differences in disembodied perception and emotional healing effects among different landscape types in Study 2.

Data collection and cleaning

Network comments about the Erhai Lake Scenic Area were collected from three websites using the Octopus Cloud Service Cluster combined with Python. The data collection and scraping tasks are mainly carried out through the Octopus service platform. After setting up the program on this platform, data can be automatically crawled. Python is mainly used to clean and process the crawled data, such as removing HTML tags, spaces, line breaks and other irrelevant information. The data collection process fully complies with the experimental terms and user agreements of each website. The above-mentioned apps selected for the research are all legal and legitimate, and the online comments within them are all open-source data voluntarily and publicly posted by users. Therefore, the online comments are publicly accessible and the data can be obtained. From January 2019 to January 2025, a total of 9,347 comments were collected. The collected data was filtered and cleaned by removing irrelevant comments that were not related to the scenic spots or themes, those with obvious marketing or advertising content, non-text comments (such as symbols, images, etc.), and duplicate comments. In the end, 8,072 valid comments were obtained.

Text data semantic analysis

After the data was collected and cleaned, this study utilized the Rost Content Mining System (Rost CM) to perform an initial semantic analysis of the comments⁵⁹. After conducting tokenization with Rost CM, word frequency analysis was conducted to extract high-frequency words. The high-frequency word results were then analyzed using social network and semantic network analysis to build a network graph, further revealing the inherent relationships and structures within the comments.

Grounded theory coding and classification

This study, based on the methodology of Grounded Theory (GT)⁶¹, utilized NVivo 12 qualitative analysis software for coding analysis, in conjunction with the results from text semantic analysis. Following the principles of GT, the study employed a combination of manual and automatic coding methods, reading the texts line by line and performing three stages of coding: Open Coding, Axial Coding, and Selective Coding. With reference to the three-level classification logic, the text content was refined into three levels: embodied perception landscape types (first level), landscape sequence elements (second level), and specific landscape components (third level). In Open Coding, researchers extracted initial concepts and summarized the potential themes of tourists’ original perceptual evaluations. Subsequently, the initial concepts were combined with high-frequency words and social semantic networks for the classification and integration in Axial Coding. Finally, core categories-landscape types were extracted and the three-level classification was detailed. The details of the coding process are shown in the figure, including statistics on reference points of coded files and a hierarchical chart of coding nodes (Fig. 3).To ensure the reliability and validity of the coding process, this study employed multiple coders in combination with theoretical review and validation⁵⁹.

Fig. 3.

Encoding node diagram (self-drawn by the author).

Sentiment analysis by landscape type

Based on the coding results, relevant online review texts corresponding to each landscape type were integrated and processed, and sentiment polarity analysis was conducted for each type using Rost CM. A horizontal comparison of emotional responses across the different landscape types was performed, along with a vertical comparison between these emotional responses and the perceived restorative strength of disembodied experiences identified in Study 2.

Results of study 1

Network text semantic analysis

Semantic networks are an application of complex network technology in the field of natural language processing. The frequency of word usage is closely related to the visitor’s experience and plays an important role in forming a sense of place⁶². Using Rost CM to screen the top 50 high-frequency words from online reviews of the Erhai Scenic Area, irrelevant words related to travel experience, landscape perception, and emotions were removed. The main focus of visitors’ embodied travel experiences in Erhai is on “scenery, photography, ancient town, beautiful, local customs, sun” etc. (Fig. 4 left). Based on the frequency results, a social and semantic network map of the high-frequency words was constructed and divided into three levels, revealing the underlying social relationship characteristics in the reviews. Among them, “scenic spots, Bai nationality, islands, nature, ancient town, guesthouses” formed the core group of high-frequency word nodes, with related words like “Cangshan, romantic, local customs, sunshine,” forming a secondary word circle (Fig. 4 right).

Fig. 4.

Network text analysis (left = word frequency, right = semantic network diagram).

Landscape perception coding classification model

The core level represents embodied perception landscape types, coded according to the overall landscape functional attributes and spatial distribution sequence: offshore ecological-natural landscape, shore-side commercial-leisure landscape, coastal historical-cultural landscape, and shoreline rural-pastoral landscape. The second level includes landscape sequence elements, categorized based on recognizable spatial elements and decoding sequences, which consist of 12 items: religious cultural landscapes, fishing village dock landscapes, and natural lake islands, etc. The third level consists of specific landscape components, totaling 21 items (Fig. 5).

Fig. 5.

Landscape classification code distribution diagram (Self-drawn by the author).

By coding and classifying embodied landscape perceptions, this study divides tourists’ landscape perceptions of the Erhai Scenic Area into four categories. To compare the healing differences and specific transformation effects among landscape types, based on the existing hypotheses H1 and H3 combined with the classification results, the hypothetical model framework of this study is refined, and the following hypotheses are proposed (Fig. 6). Among them, H1a-d verify the differences of these four types of landscapes in terms of emotional changes, aiming to reveal which landscape type can more effectively achieve digital healing. H3a-d verify the differentiated driving forces of the four landscape types on travel intentions, so as to explore the specific transformation effects of different landscapes from disembodied perception to embodied travel intentions. The verification results of the hypotheses will fill the research gap in the absence of a mechanism for differentiating landscape types, provide precise guidance for short video content creation and destination marketing, and maximize health benefits and commercial transformation.

Fig. 6.

Hypothesis conceptual model (self-drawn by the author).

H1a

Disembodied perception of offshore ecological-natural landscapes has a positive effect on emotional changes.

H1b

Disembodied perception of shore-side commercial-leisure landscapes has a positive effect on emotional changes.

H1c

Disembodied perception of coastal historical-cultural landscapes has a positive effect on emotional changes.

H1d

Disembodied perception of shoreline rural-pastoral landscapes has a positive effect on emotional changes.

H3a

Disembodied perception of offshore ecological-natural landscapes has a positive effect on travel intention.

H3b

Disembodied perception of shore-side commercial-leisure landscapes has a positive effect on travel intention.

H3c

Disembodied perception of coastal historical-cultural has a positive effect on travel intention.

H3d

Disembodied perception of shoreline rural-pastoral landscapes has a positive effect on travel intention.

Emotional analysis of the four embodied perception landscape types

Based on the results of coding and classification, sentiment analysis was conducted on evaluative texts corresponding to the four core landscape types (Table 1). The findings show that, following embodied experiences, all four landscape types elicited predominantly positive emotional responses from visitors. Among them, historical-cultural and rural-pastoral landscapes generated the highest levels of emotional positivity, followed by ecological-natural landscapes. In contrast, commercial-leisure landscapes exhibited the highest level of negative sentiment.

Table 1.

Emotional analysis of the four perception landscape types.

		Ecological-natural landscape	Commercial-leisure landscape	Historical-cultural landscape	Rural-pastoral landscape
Positive emotions		82.63%	83.27%	86.66%	85.14%
Positive emotions segmentation	Low	22.75%	16.91%	20.49%	21.90%
	Moderate	21.06%	24.86%	23.73%	23.53%
	High	38.82%	41.50%	42.45%	39.71%
Neutral emotions		5.46%	3.16%	3.76%	4.49%
Negative emotions		11.91%	13.56%	9.58%	10.37%
Negative emotions segment	Low	7.69%	9.22%	6.56%	7.82%
	Moderate	3.00%	2.26%	1.62%	1.70%
	High	0.54%	0.99%	0.66%	0.39%

Study 2

Methods of study 2

Questionnaire design

Building upon the findings of Study 1, Study 2 was designed to examine the effect of disembodied 2D travel videos on emotional changes and embodied travel intention, as well as to explore the relationships among these three variables. Additionally, the study aimed to assess whether two personality traits—openness and neuroticism—moderate the relationship between landscape perception and emotional changes. In accordance with the ethical review requirements of Central South University, this study does not involve life science or medical research and thus does not require ethical approval. However, informed consent must be obtained from all participants and their legal guardians during the survey. An informed consent form containing relevant ethical requirements is placed at the beginning of the questionnaire, and only participants who provide their consent are permitted to proceed with the subsequent survey. In addition, this study has been reported to the Academic Committee of the School of Architecture and Art at Central South University, and the entire survey was strictly conducted in accordance with the relevant requirements under the supervision of the supervisor. This study used a survey questionnaire method, which includes four sub-scales: the Big Five Personality Scale⁶³, the Emotional Change Scale⁶⁴, the Sensory Perception Dimension Scale⁶⁵, and the Travel Intention Scale⁶⁶ and a video segment (https://www.bilibili.com/video/BV1xrfhYpELX/?spm_id_from=333.1387.homepage.video_card.click). In the process of questionnaire distribution, to facilitate participants' completion, all Chinese translation versions employed have undergone validation and been published in academic literature^67–69. These scales measure nine variables with a total of 59 items. Following questions about demographic characteristics, the first set assessed participants’ levels of openness and neuroticism using the eight and ten items respectively from the Big Five Inventory⁶³. The items were rated on a 5-point Likert scale, ranging from 1 (“disagree strongly”) to 5 (“agree strongly”). Then comes the experimental part, participants were asked to watch a 5-min video. To ensure that each participant carefully watches the videos, we implemented a page redirect mechanism at the video link section in the questionnaire. Participants must click the video link, jump to the video webpage, and complete the viewing before they can proceed to the next step of questionnaire filling. Meanwhile, the number of viewers and views will be tracked through the video platform’s backend statistics. The video was divided into two distinct sections, the first section lasts for 90 s and depicts the daily work, study, and life of individuals from various professions and age groups, helping the participants recall their daily routines. Following this, there is a 20-s subtitle prompt encouraging participants to reflect on and recall their emotional state throughout the day. Then, there is a 10-s pause, allowing participants to return to their regular emotional state. The second section features a 3-min travel video of Erhai Lake, showcasing four types of landscapes in sequence: offshore natural landscapes, shore-side commercial landscapes, coastal cultural landscapes, and shoreline pastoral landscapes. Each landscape type is displayed for 45 s. After viewing the video, participants answered six questions aimed at assessing emotional changes before and after watching the travel video, based on the scale developed by Rogers et al⁶⁴. This 6-descriptor scale, derived from the Positive and Negative Affect Schedule (PANAS), includes four negative emotions (nervous, irritable, upset, distressed) and two positive emotions (excited, and happy). The original 5-point scale was modified to assess emotion variance, with the following scale: 1 = much less than before, 2 = less than before, 3 = about the same, 4 = more than before, 5 = much more than before⁶⁴. Subsequently, four sets of questions were presented to evaluate participants’ perception of four landscape types. Referencing Li’s study⁶⁵, each set included eight dimensions, with perception intensity rated after each dimension. Finally, participants were asked to complete the travel intention scale based on their feelings after watching the video. This marks the completion of the entire experimental procedure⁶⁶.

In the design of the travel videos, several approaches were attempted to avoid mutual influence between different landscape types. Initially, the four types of landscape videos were presented separately, with different participants viewing distinct categories of landscapes. However, the results were significantly affected by individual differences among participants. We also experimented with randomizing the playback order of the four videos, yet this approach still failed to prevent inter-video influence and comparison, while also leading to sequence confusion in participants’ questionnaire responses. After discussions within our team, we recognized that while the comparison between participants might enhance the differentiation of the landscape types, this form of differentiated experimental result better reflects the data and achieves the experimental goal by emphasizing these differences. Moreover, we considered that landscape perception constitutes a holistic experience of the environment, often arising from the interplay of various landscape elements. Therefore, imposing strict controls over inter-landscape influences could disrupt the natural experiential perception of the participants. Additionally, we implemented baseline emotional calibration and visual variable control to minimize other interfering factors. Before each viewing, participants were given 90 s to recall daily life scenarios, followed by 20 s of subtitle guidance and a 10-s black screen pause to help return their emotions to baseline. We also standardized the production of the four types of landscape videos to ensure uniformity in duration, resolution, color saturation, and background music.

Data collection

During the questionnaire distribution process, a combined method of online and offline distribution was employed. An online questionnaire was created using the Wenjuanxing platform. Initial data collection was conducted through online and offline convenience sampling from 27, January to 1st, February, 2025. Online questionnaires were filled out through WeChat forwarding and group messaging, while in terms of offline, researchers went to communities and surrounding public places to find volunteers to help fill it out. During the offline data collection, since some participants were unwilling to scan the QR code or had difficulty opening the video, electronic devices were provided. 95 questionnaires were completed in the pre-experiment, with 94 valid ones. The reliability and validity analysis of the data met the standards, with a KMO value of 0.790 and a Cronbach’s α value of 0.955, indicating good reliability and validity in the pre-experiment (See supplementary material).The second-stage collection used a snowballing method from 1st, February to 7th, February, 2025, resulting in a total of 462 completed questionnaires, of which 409 responses were valid, with an effective rate of 88.5%. Participants were mostly from Mainland China, representing a diverse range of careers and age groups. The participants were consisted of 49.1% male participants (n = 201) and 50.9% female participants (n = 208). As shown in Table 2, the sample was dominated by people aged 18 to 50, accounting for 83.3%. This study primarily focuses on alleviating emotions through travel-scenery videos within virtual networks in the digital age. Individuals aged 18 to 50 constitute both the main force of the internet era and the demographic facing the greatest pressures in daily life. Therefore, the sample population was highly consistent the primary target audience of this research⁵⁸.

Table 2.

Population statistics.

Variable	Category	Frequency	Percentage
Profile of the sample (n = 409)
Gender	Male	201	49.10%
	Female	208	50.90%
Age	< 18	32	7.80%
	18–25	114	27.90%
	26–30	93	22.70%
	31–40	70	17.10%
	41–50	64	15.60%
	51–60	26	6.40%
	> 60	10	2.40%
Education	Junior high school or below	54	13.20%
	Senior high school or vocational school	83	20.30%
	Higher vocational collage	99	24.20%
	Bachelor’s degree	114	27.90%
	Master’s degree or above	59	14.40%
Occupation	Agencies/enterprises employee	130	31.80%
	Self-employed	50	12.20%
	Labourer	16	3.90%
	Service/Salesperson	50	12.20%
	Professional/educational technical personnel	48	11.70%
	Student	81	19.80%
	Housewife/husband	16	3.90%
	Other	18	4.40%
Travel Frequency	0/year	44	10.80%
	1–2/year	233	57%
	3–4/year	101	24.70%
	> 5/year	31	7.60%

Results of study 2

Reliability and validity test

SPSS 27.0 was used in data analysis, and all scales in the study were maturity scales used and cited by previous researchers. As shown in Table 3, the Cronbach’s α value of Neuroticism and Emotional changes scales were all > 0.60, which was acceptable, while the rest ranged between 0.788 to 0.937 for all variables, which were all > 0.70, indicating good reliaiblity. The composite reliability (CR) and were between 0.896 and 0.930, and the standardized factor loading and average variance extracted (AVE) values of all items were also calculated to assess the reliability of the questionnaire. All standardized factor loading and AVE values exceeded the 0.50, and the all CR values were greater than 0.70, indicating that both reliability and convergent validity were acceptable. Finally, KMO and Bartlett’s test are used to assess the validity of data, the KMO value was above 0.90, and the Bartlett’s test result was significant, indicating marvelous validity.

Table 3.

Validity and reliability test.

Validity test
KMO measure of sampling adequecy		0.927
Bartlett’s test of sphericity	Approx. Chi-Square	12,464.408
	df	1711
	sig	0.000

Reliability test
Latent Variable	Item	Cronbach’s α	Standardized factor loading	CR	AVE
Total		0.937
Neuroticism (NE)	NE1: I see myself as someone who is depressed, blue	0.611	0.749	0.922	0.597
	NE2 (R): I see myself as someone who is relaxed, handles stress well		0.882
	NE3: I see myself as someone who can be tense		0.773
	NE4: I see myself as someone who worries a lot		0.775
	NE5 (R): I see myself as someone who is emotionally stable, not easily upset		0.804
	NE6: I see myself as someone who can be moody		0.702
	NE7 (R): I see myself as someone who remains calm in tense situations		0.717
	NE8: I see myself as someone who gets nervous easily		0.764
Openness (OP)	OP1: I see myself as someone who is original, comes up with new ideas	0.788	0.827	0.920	0.541
	OP2: I see myself as someone who is curious about many different things		0.794
	OP3: I see myself as someone who is ingeneous, a deep thinker		0.754
	OP4: I see myself as someone who has an active imagination		0.721
	OP5: I see myself as someone who is inventive		0.739
	OP6: I see myself as someone who values artistic, aesthetic experiences		0.588
	OP7 (R): I see myself as someone who prefers work that is routine		0.585
	OP8: I see myself as someone who likes to reflect, play with ideas		0.804
	OP9 (R): I see myself as someone who has few artistic interests		0.903
	OP10: I see myself as someone who is sophisticated in art, music, or literature		0.561
Emotional changes	Nervous	0.681	0.680	0.930	0.691
	Upset		0.827
	Excited		0.848
	Irritable		0.856
	Distressed		0.878
	Happy		0.882
PSD for offshore ecological- natural landscape (PSD1)	Social: Here is an environment suitable for social activities	0.863	0.724	0.943	0.676
	Space: This is a spacious and undisturbed environment		0.804
	Nature: The sensation of wilderness and nature		0.810
	Refuge: Here is an enclosed and safe environment		0.865
	Prospect: Here is an open space with a wide view		0.828
	Serene: Here is a silent and peaceful environment		0.791
	Culture: Many artificial elements are decorated here		0.925
	Rich in species: many animals ang plants are around here		0.817
PSD for shore-side commercial-leisure landscape (PSD2)	Social: Here is an environment suitable for social activities	0.874	0.916	0.896	0.529
	Space: This is a spacious and undisturbed environment		0.640
	Nature: The sensation of wilderness and nature		0.720
	Refuge: Here is an enclosed and safe environment		0.820
	Prospect: Here is an open space with a wide view		0.596
	Serene: Here is a silent and peaceful environment		0.572
	Culture: Many artificial elements are decorated here		0.917
	Rich in species: many animals ang plants are around here		0.519
PSD for historical-cultural landscape (PSD3)	Social: Here is an environment suitable for social activities	0.873	0.788	0.905	0.551
	Space: This is a spacious and undisturbed environment		0.620
	Nature: The sensation of wilderness and nature		0.804
	Refuge: Here is an enclosed and safe environment		0.799
	Prospect: Here is an open space with a wide view		0.550
	Serene: Here is a silent and peaceful environment		0.569
	Culture: Many artificial elements are decorated here		0.910
	Rich in species: many animals ang plants are around here		0.816
PSD for rural-pastoral landscape (PSD4)	Social: Here is an environment suitable for social activities	0.865	0.728	0.902	0.541
	Space: This is a spacious and undisturbed environment		0.550
	Nature: The sensation of wilderness and nature		0.874
	Refuge: Here is an enclosed and safe environment		0.787
	Prospect: Here is an open space with a wide view		0.573
	Serene: Here is a silent and peaceful environment		0.623
	Culture: Many artificial elements are decorated here		0.925
	Rich in species: many animals ang plants are around here		0.766
Ttravel intention (TI)	TI1: Given the opportunity, I will visit locations I have seen in the video	0.851	0.895	0.925	0.805
	TI2: I will speak positively of the video’s destinations		0.889
	TI3: I recommend the trip to my relatives and friends to the video recording destination		0.908

Correlation analysis

In order to test whether the hypotheses 1, 3, and 4 were supported, correlation tests were first conducted between all variables, excluding the personality traits. As shown in Table 4, a Pearson correlation analysis was performed to examine the relationships between landscape perception (total and four types), emotional changes, and travel intention scores. The analysis revealed significant positive correlations for each pair of variables, with Pearson’s r values ranging from 0.354 to 0.820. These results suggest moderate to strong relationships between the variables, indicating that further regression analysis can be conducted.

Table 4.

Correlation analysis.

	Zero-order correlations
	1	2	3	4	5	6	7
1. Total landscape perception	–
2. Emotional changes	0.463**	–
3. Travel intention	0.610**	0.404**	–
4. Ecological-natural landscape perception	0.800**	0.384**	0.527**	–
5. Commercial-leisure landscape perception	0.798**	0.354**	0.407**	0.505**	–
6. Historical-cultural landscape perception	0.810**	0.382**	0.507**	0.498**	0.546**	–
7. Rural-pastoral landscape perception	0.820**	0.376**	0.530**	0.585**	0.514**	0.560**	–

***p < .001, **p < .01, *p < .05.

Main effect test

The study continued employing linear and hierarchical multiple regression analysis using SPSS 27.0, with Model 1 and Model 2, to test the Hypotheses 1, 3, 4, along with their corresponding sub-hypotheses (H1a, H1b, H1c, H1d, H3a, H3b, H3c, and H3d). Respondents’ demographic variables, including gender, age, education, occupation, and travel frequency, were included as control variables in the analysis. The majority of respondents reported increases in positive affect and decreases in negative affect after watching the disembodied travel video, where the response counts and histograms for each item in the Mood Change Scale are shown in Table 5. As shown in Table 6, the results of the main effect indicated that overall landscape perception had a significant positive effect on both emotional changes (β = 0.412, p < 0.001) and embodied travel intention (β = 0.591, p < 0.001). Additionally, emotional changes were found to significantly predict travel intention (β = 0.378, p < 0.001). These findings supported Hypotheses 1, 3, and 4. Further analysis using multiple regression revealed that, with the exception of the rural landscape, all other landscape types had a significant positive effect on emotional changes. Similarly, all landscape types, except for shore-side commercial landscapes, were positively associated with travel intention. These results provide support for H1a, H1b, H1c, H3a, H3c, and H3d. H1d and H3b were not supported.

Table 5.

Emotional change scale & histogram

Table 6.

Main effects testing results.

Path	B	SE	β	t	p	LLCI	ULCI	Support
H1 Landscape perception – > Emotional changes	0.397	0.420	0.412	9.395	< .001	0.314	0.480	Yes
H1a Ecological-natural landscape – > Emotional changes	0.109	0.045	0.138	2.394	< .05			Yes
H1b Commercial-leisure landscape – > Emotional changes	0.091	0.042	0.119	2.151	< .05			Yes
H1c Historical-cultural landscape – > Emotional changes	0.138	0.043	0.181	3.212	0.001			Yes
H1d Rural-pastoral landscape – > Emotional changes	0.056	0.047	0.071	1.196	0.232			No
H3 Landscape perception – > Travel intention	0.896	0.061	0.591	14.655	< .001	0.776	1.016	Yes
H3a Ecological-natural landscape – > Travel intention	0.306	0.065	0.248	4.715	< .001			Yes
H3b Commercial-leisure landscape – > Travel intention	0.029	0.060	0.024	0.482	0.630			No
H3c Historical-cultural landscape – > Travel intention	0.285	0.061	0.238	4.639	< .001			Yes
H3d Rural-pastoral landscape – > Travel intention	0.282	0.068	0.224	4.175	< .001			Yes
H4 Emotional changes – > Travel intention	0.596	0.075	0.378	7.897	< .001	0.447	0.744	Yes

Analysis of moderating effect

The study used PROCESS extension v4.0 (Model 1) to conduct hierarchical regression analysis to test the moderating effects of personality factors (neuroticism and openness) on the relationship between landscape perception (X) and emotional change (Y), while controlling the demographic variables of gender, age, occupation, education, and travel frequency (Fig. 7). As shown in Table 7, the results showed that neuroticism and openness significantly moderated the effect of X on Y, supporting H2a and H2b.

Fig. 7.

Moderating effect results (left = neuroticism, right = openness).

Table 7.

Moderating and mediating effect results.

Outcome variables	Predictor variables	R	R2	F	B	β	t	p	LLCI	ULCI
Moderating effect results
Emotional changes		0.554	0.307	22.142***
	Gender				− 0.087	− 0.066	− 1.550	0.122	− 0.198	0.023
	Age				0.071	0.161	3.570***	0.000	0.032	0.111
	Education				0.057	0.195	4.451***	0.000	0.032	0.082
	Occupation				0.055	0.104	2.295*	0.022	0.008	0.102
	Travel Frequency				− 0.035	− 0.040	− 0.930	0.353	− 0.109	0.039
	Landscape perception				0.365	0.379	8.556***	0.000	0.281	0.448
	Neuroticism				− 0.026	− 0.022	− 0.481	0.631	− 0.131	0.080
	Int_1				0.316	0.202	4.607***	0.000	0.181	0.450
Emotional changes		0.590	0.348	26.636***
	Gender				− 0.061	− 0.046	− 1.111	0.267	− 0.168	0.047
	Age				0.077	0.173	3.966***	0.000	0.039	0.115
	Education				0.051	0.176	4.122***	0.000	0.027	0.076
	Occupation				0.051	0.096	2.175*	0.030	0.005	0.097
	Travel Frequency				− 0.052	− 0.059	− 1.429	0.154	− 0.124	0.020
	Landscape perception				0.320	0.332	7.168***	0.000	0.232	0.407
	Openness				0.099	0.098	2.092*	0.037	0.006	0.193
	Int_1				0.357	0.259	6.258***	0.000	0.245	0.469

Path	B	SE	β	t	p	LLCI	ULCI	Support
Mediating effect results
X– > M	0.397	0.042	0.412	9.395	< .001	0.314	0.480	Yes
X– > Y	0.805	0.067	0.531	12.058	< .001	0.674	0.937	Yes
M– > Y	0.230	0.071	0.146	3.220	0.001	0.090	0.371	Yes
Total effect of X on Y	0.896	0.061	0.591	14.655	< .001	0.776	1.016
Direct effect of X on Y	0.805	0.067		12.058	< .001	0.674	0.936
Indirect effect of X on Y	0.091	0.029				0.038	0.152

The result revealed a significant interaction between X and openness (B = 0.357, p < 0.001, 95% CI [0.245, 0.469]), with the explanatory power of R² = 0.348, indicating that the relationship between X and Y was moderated by openness level. Specifically, when openness was at a higher level (+ 1SD), the relationship between X and Y was stronger (B = 0.554, p < 0.001), suggesting that openness enhances the effect of landscape perception. However, lower opennes level (-1SD) had no significant moderating effect (p = 0.157). The regression equation is expressed as:

The neuroticism moderation model was significant overall (F = 22.142, p < 0.001), with an explanatory power of R² = 0.307. The interaction term between neuroticism and landscape perception significantly predicted emotional changes (B = 0.316, p < 0.001, 95% CI [0.181, 0.450]), indicating that neuroticism significantly positively moderates the relationship between landscape perception and emotional changes. Specifically, when neuroticism was lower (-1SD), the moderating effect exists (B = 0.188, p < 0.01). When neuroticism was at a higher level (+ 1SD), the relationship between X and Y was stronger (B = 0.554, p < 0.001). The regression equation is expressed as:

Analysis of mediating effect

As shown in Table 7, after controlling for gender, age, occupation, education level, and travel frequency, the results showed that emotional change plays a significant partial mediating role in the relationship between landscape perception and travel intention. The total effect of landscape perception on travel intention was 0.896 (p < 0.001), with the direct effect being 0.805 (p < 0.001) and the indirect effect being 0.091 (95% CI [0.038, 0.152]), which accounts for 10.2% of the total effect. The partial mediation effect was significant. Therefore, Hypothesis H5 is supported. As shown in the path analysis diagram (Fig. 8), this study reports the results of all hypothesis tests. Except for the non-significant effects of the coastal commercial landscape on travel intention and the shoreland pastoral landscape on emotional changes, all the proposed paths in the model show significant positive correlations.

Fig. 8.

Path model of the effects (self-drawn by the author).

Comparison of study 1 and study 2 results

A comparative analysis was conducted between the emotional responses elicited by embodied experiences of the four landscape types in Study 1 and the perceived restorative strength in disembodied contexts in Study 2 (Table 8). This analysis included both a horizontal comparison across the four landscape types and a vertical comparison between the degree of disembodied perception and the valence of emotions following embodied experiences, aiming to capture the emotional shifts from virtual viewing to on-site experience. To more intuitively compare the degree of embodied versus disembodied perception, we represent it using a numerical model:

Table 8.

Comparison of embodied experiences and disembodied perception.

		Ecological-natural landscape	Commercial-leisure landscape	Historical-cultural landscape	Rural-pastoral landscape
Embodied experiences	Positive (SP)	82.63%	83.27%	86.66%	85.14%
	Neutral (SO)	5.46%	3.16%	3.76%	4.49%
	Negative (SN)	11.91%	13.56%	9.58%	10.37%
	ME	0.7072	0.6971	0.7708	0.7477
Disembodied perception	Strong (SS)	60.82%	48.27%	58.23%	60.61%
	Neutral (SN)	4.33%	6.06%	6.71%	5.19%
	Weak (SW)	34.85%	45.67%	35.06%	34.20%
	MD	0.2597	0.0260	0.2317	0.2641

In this model, a, b, and c represent the coefficients for positive, neutral, and negative emotions, respectively, assigned values of 1, 0, and -1. Horizontal analysis reveals that in embodied emotional perception: historical-cultural > rural-pastoral > ecological-natural > commercial-leisure (M_E:0.771 > 0.748 > 0.707 > 0.697); in the degree of disembodied perception: rural-pastoral > ecological-natural > historical-cultural > commercial-leisure (M_D:0.264 > 0.260 > 0.232 > 0.026). Vertical analysis shows that during the transition from disembodied to embodied perception, the change value is: commercial-leisure > historical-cultural > rural-pastoral > ecological-natural.

Discussion

Against the backdrop of digitalization and the rise of “cloud tourism,” disembodied travel has emerged as a new form of tourism. To uncover the mechanism of “perception–emotion–behavior” in disembodied landscape experiences and the differentiated effects of landscape types, this study adopts a combination of quantitative and qualitative methods. A qualitative analysis of online text data was conducted using grounded theory, followed by video experiments and questionnaire surveys based on the coding results. The survey data were then subjected to statistical analysis. Findings aim to maximize the health benefits and commercial potential of “cloud tourism”, establishing a new paradigm for health tourism research in the digital intelligence era.

Theoretical contribution: extending the landscape—emotion link into the virtual context

Social media enables emotional connection and healing through virtual experiences. This study establishes a mediating mechanism of “disembodied landscape perception → embodied emotional change → embodied travel intention (B = 0.896, p < 0.001),” supporting environmental restorative theory, attention restoration theory(ART), stress recovery theory(SRT), and human–nature connectedness theory(HNC)^70,71. It extends Embodied Cognition Theory(ECT)¹ into virtual contexts by introducing environmental psychology and HNC, while proposing a bidirectional “disembodied-embodied” transformation model. The results confirm that disembodied perception can evoke embodied-like emotional responses⁷, consistent with findings on digital nature’s effectiveness in stress reduction^72,73. Visual nature representations such as images and videos were shown to alleviate anxiety and benefit mental health^74,75. Additionally, the study aligns with existing research indicating that positive short video experiences strengthen travel motivation^76,77 and support tourism promotion⁷⁸. The stimulus-organism-response(SOR)⁷⁹ framework is also extended to social media’s disembodied context⁸⁰. These findings highlight that although landscape videos offer healing potential, they cannot replace real-world interaction. Instead, they should function as mediators encouraging physical travel. Finally, the study underscores personality’s moderating role in landscape healing (Openness: B = 0.357, p < 0.001, Neuroticism: B = 0.316, p < 0.001), establishing a personalized healing paradigm^81–83. Questionnaire results validate neuroticism’s bidirectional moderation and openness’s threshold effect, supporting the notion that “the same landscape may be perceived differently” based on individual traits⁸⁴.

Analysis of landscape healing and transformative effects differences

The findings address the research gap concerning differential restorative effects across landscape types^85,86. Through multivariate regression analysis of four primary landscapes in the Erhai Scenic Area, results indicate that all types positively influenced emotions or travel intentions to varying degrees⁸⁷, supporting the applicability of environmental restorative theory in digital contexts²² (p < 0.05, H1abc&H3acd were supported). However, rural-pastoral landscape had no significant impact on emotional changes(p = 0.232, H1d was not supported), disembodied perceptions of farmland, docks, villages, agricultural labor, and fishing scenes do not effectively alleviate emotions. Commercial-leisure landscapes can produce a psychological healing effect ,but can’t significantly influence travel intention and further stimulate travel intentions and real-world tourism consumption(p = 0.630, H3b was not supported). Comparative data from Studies 1 and 2 indicate landscape-specific effects in the transition from embodied to disembodied perception. Ecological-natural landscapes show high perceptual accessibility but limited emotional shift (M_E = 0.707, M_D = 0.260), supporting “digital nature therapy.” Historical-cultural landscapes achieve high conversion, relying strongly on embodied on-site interaction (M_E = 0.771, M_D = 0.232). Commercial-leisure landscapes are weak in both perceptions, reaffirming the limitations of standardized environments, yet their high emotional transition rate suggests potential within the “check-in economy.”(M_E = 0.697, M_D = 0.026) A notable paradox was observed in rural-pastoral landscapes, which performed well perceptually but elicited weak emotional response (M_E = 0.747, M_D = 0.264, H1d was not supported). While empirical data cannot fully explain this discrepancy, we attempted to speculate on the reasons: this paradox may stem from the social reality of rural areas. They are often perceived as lacking convenient transportation, adequate services, education, and employment opportunities, prompting many young people to leave^88,89. As a result, rural youth experience a conflicted relationship with their hometowns, marked by both attachment and detachment, pride and entrapment^90,91. While urban youth often lack emotional connections to the countryside altogether. This tension may help explain the observed contradictions.

Practical implications for health tourism under social media

Future health tourism are likely to integrate both disembodied and embodied experiences, forming a closed loop of “embodied–disembodied” travel. While social media can provide psychological comfort and serve as a supplementary tool for mental health, it cannot substitute for the effects of on-site travel⁵¹. Based on the new tourism model constructed in this study (Fig. 9), a novel model centered on the “digital-physical” value transformation closed loop, the core principle of “precision healing” is established at the conceptual framework level. Its core dimensions adhere to the specificity of landscape types. Drawing on empirical results, it distinguishes the distinct emotional restoration efficacy and behavioral transformation pathways of the four landscape types across the two dimensions of “disembodied perception” and “embodied experience.” Furthermore, the healing effects are moderated by the user’s personality traits, identifying “Neuroticism” and “Openness” as key moderating variables and proposing the “personality-landscape matching” principle. At the path mechanism level, the model optimizes the “digital-physical” conversion process. Targeting the “disembodied perception → emotional change → travel intention” pathway, short videos serve as the initial touchpoint, providing immediate psychological solace by triggering a “sense of pseudo-presence.” For digital front-end outreach, content platforms and creators should leverage algorithms for personalized content recommendations. For instance, users high in Neuroticism could be directed towards ecological natural landscapes to establish emotional safety, while users high in Openness could be offered historical and cultural landscapes to stimulate cognitive interest, thereby maximizing the efficiency of front-end emotional arousal. At the critical conversion node, destination marketers should allocate resources based on the unique conversion characteristics of each landscape type. For example, focus on developing in-depth experience products for historical and cultural landscapes, which show high conversion rates, while for commercial leisure landscapes, the marketing emphasis should be placed on their immediate psychological solace value. Finally, the embodied experience of physical travel feeds back into the digital ecosystem through user-generated content sharing, enhancing the model’s sustainability and creating a value closed loop of “digital preview—physical experience—social feedback.” At the level of managerial implications, content creators should integrate the creative principles of health and healing into short video production, evoking user emotional resonance and psychological relaxation. This involves prioritizing the integration of restorative content such as natural, cultural, and commercial landscapes. Destination marketers should reduce reliance on homogenized commercial landscapes. By translating inspiration into concrete action, such as launching short-term “Weekend Healing Getaways,” they can lower the time and economic thresholds for participation. This guides a shift from transient online digital healing to sustained real-world ecological restoration, which can not only help prevent social media addiction but also construct a virtuous cycle for the healthy tourism model. For the overall industry, there is a need for collective vigilance against the “traffic-first” algorithm tendency, fostering a more positive digital environment, and guiding the industry away from fleeting sensory stimulation towards a new healthy tourism model offering sustained therapeutic value.

Fig. 9.

Schematic diagram of new health tourism (self-drawn by the author).

Implicit ethical challenges of disembodied digital tourism

The psychedelic digital space, chaotic space proliferation, ambiguous spatial boundaries, and poor design ethics could threaten human safety and undermine human dominance as space subjects. Humans are influenced by the spatial–temporal dynamics of information and virtual spaces, blurring the boundaries between embodied and disembodied spaces. This study highlights concerns over implicit digital ethics in the age of media-driven tourism. In the new health tourism, it is necessary to establish a clearer boundary between disembodied and embodied space, and optimize the immersion of disembodied experience without replacing the real experience. At the technical level, a “reality anchoring system” could be developed by setting thresholds for digital immersion (e.g., 60 min) and integrating biofeedback sensors such as heart-rate and eye-tracking to monitor cognitive load, thereby dynamically adjusting information flow to prevent overload and imbalance. Within the “influencer economy,” it is equally crucial to establish a responsible communication ecosystem that avoids excessive filters and misleading promotions, which distort user perceptions and trigger herd effects. Instead, algorithms should be guided to channel traffic in ways that encourage ecological protection of destinations. Ultimately, addressing these challenges requires sustained collaboration among governments, researchers, content creators, and users. By aligning technological innovation with ethical governance, a balanced and symbiotic relationship between “digital twins” and the real world can be achieved.

Limitations and future research directions

This study has several limitations. First, there may be differences in perception based on age and generational factors. Since this study primarily focuses on younger and middle-aged groups, individuals under 40 years old occupied 75.5% of the total sample. Therefore, the sample group may not experience the same level of embodied emotional connection with pastoral landscapes influenced by short video “filters” as older adults do, which leads to the non-significant effect of rural-pastoral landscape on emotional change (H1d is not supported). Furthermore, there are also some potential factors that may have influenced participants’ responses to the video stimuli. For instance, participants’ cross-comparisons of different landscape videos and variations in playback order could have introduced biases into the experimental results. Future studies should address and refine these design flaws. In addition, the duration of the video can cause changes in the interaction threshold. Research indicates that short videos (< 1 min) are more likely to trigger emotional peaks but with poor persistence, while long videos (> 5 min) promote deeper cognition but may lead to aesthetic fatigue⁹². This study used a 3-min landscape video for the experiment, and future research could establish a “duration-emotion curve” model to optimize the matching of content duration with landscape types. Moreover, subsequent experiments could further explore the healing effects of various disembodied perceptual elements, such as media formats (e.g., images, videos, VR), background music, and filming techniques.

Conclusion

This study aims to explore the digital healing mechanisms of disembodied landscape perception through social media, focusing on differences in restorative effects across landscape types and their specific transformation effects. The findings reveal: (1) The healing mechanism of “disembodied landscape perception—embodied emotional change—embodied travel intention” was validated, showing emotional change as a partial mediator between perception and travel intention, with personality traits (openness and neuroticism) exerting moderating effects; (2) All four landscape types positively influenced either emotions or travel intention to varying degrees, only rural-pastoral landscapes showing no significant impact on emotional change and commercial-leisure landscapes showing no significant effect on travel intention; (3) When comparing the strength of disembodied perception with the positive and negative emotions of embodied experiences, commercial-leisure and historical-cultural landscapes exhibited the highest transformation rates, whereas rural-pastoral and ecological-natural landscapes showed relatively smaller changes. This study highlights the influence of social media travel videos on mental health, extends the theoretical boundaries of “landscape restoration,” and confirms the psychological healing potential of disembodied travel perception. It further advances the dual evolution of “Internet traffic economy” and “health healing” in “cloud tourism,” contributing to a new model of health tourism that integrates disembodied and embodied experiences in the digital-intelligence era.

Author contributions

Conceptualization, R.G. and Y.Q.; methodology, R.G.; software, R.G.; validation, Y.Q. and R.L.; formal analysis, Y.Q.; investigation, X.X.; resources, X.X.; data curation, R.L.; writing—original draft preparation, R.G.; writing—review and editing, Y.Q.; visualization, R.L.; supervision, R.L.; project administration, X.X.; funding acquisition, X.X. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the “High‑End Think Tank” Project of the Central South University: Study on Rural Space Design and Development Strategy under the New Development Philosophy [grant numbers 2022znzk12]; and the National Local Joint Engineering Laboratory of Digital Protection and Creative Utilization Technology for Traditional Village and Town Culture [grant number 2024HSKFJJ005].

Data availability

The data that supports the findings of this study are available in the supplementary material of this article.

Declarations

Competing interests

The authors declare no competing interests.