Emotions deteriorate gastrointestinal health: Diagnosing problems through artificial intelligence and psychometric and psycholinguistic techniques

Abstract

BACKGROUND

Psychometric tools can indicate emotions causing gastrointestinal problems while psycholinguistics analysis detects linguistic attitudes. Artificial intelligence (AI) algorithms can assess subtle changes in emotional expression. Therefore, this study was conducted to diagnose early gastrointestinal disorders like inflammatory bowel disease and irritable bowel syndrome. Language attitude, psychometric analysis, and active learning algorithms indicate that emotional expressions are a significant predictor of gastrointestinal health worsening.

AIM

To explore the consequences of emotional expression on gastrointestinal health focusing how psycholinguistics, psychologists, and AI can diagnose gastrointestinal disorders.

METHODS

This observational study used a mixed-method analyzing associations between gastrointestinal health and emotional expression. It explored psychometric assessment, language attitude, and AI-driven DxGPT analysis over the purposive sampling of 250 participants. The study used various scales. DxGPT AI analyzed the medical records and emotional stances. Quantitative data was analyzed using SmartPLS software. The qualitative data on emotional expression was studied through content analysis guided by gut-brain axis and relevance theories. These methodologies flagged at-risk profiles for further clinical assessment.

RESULTS

A significant correlation between emotional expression and gastrointestinal health was observed. The quantitative study indicated that poorer gastrointestinal health on the irritable bowel syndrome quality of life scale and higher emotional expression scores on various scales correlated with emotional patterns found by DxGPT AI driven analysis of medical records and emotional attitudes. The quality of life mediates emotional expression and gastrointestinal health, and resilience and coping moderate relations. The qualitative content analysis highlighted that recurring emotional expression is linked to gastrointestinal issues, and it was consistent with the gut-brain axis theory. The findings also supported the hypothesis that emotional issues predict gastrointestinal deterioration.

CONCLUSION

Our findings may lead to improvement of the psychological well-being of patients with gastrointestinal issues by guiding psychologists, linguists, AI professionals, and gastroenterologists to collaborate more effectively.

Key Words: Emotional expression; Gastrointestinal health; Gut-brain axis; Artificial intelligence; Irritable bowel syndrome

Core Tip: This study explored how emotional expression impacted gastrointestinal health by integrating artificial intelligence (AI), psychometric tools, and psycholinguistic techniques. Using the AI tool, DxGPT, alongside psychological and linguistic assessments, the study identified emotional markers predictive of conditions such as irritable bowel syndrome. By applying a mixed-method approach, the study supported the gut-brain axis framework and introduced a novel AI-enhanced diagnostic model. The findings highlighted the importance of emotion-based screening in early gastrointestinal disorder detection, offering interdisciplinary insights for psychologists, linguists, and gastroenterologists.

INTRODUCTION

One complication in gastroenterological health is irritable bowel syndrome (IBS). It is caused by a sensitive, nervous, tense, and worrisome temperament[1]. Strong emotions affect organs under the control of the autonomic nervous system[2]. Furthermore, an intricate association exists between physical health and emotions, concerning both the psychological and medical fields[3,4]. Gastrointestinal health is particularly vulnerable to emotional disturbances[5]. Negative emotions, anxiety, stress, and depression are linked to the onset and exacerbation of a range of gastrointestinal disorders including IBS[6]. The communication between the gut and mind is bidirectional, involving the brain-gut axis, highlighting the fact that emotions can significantly impact gastrointestinal health[7].

The advances in artificial intelligence (AI) and psychometric assessments are new dimensions that offer the potential to diagnose and treat the emotional factors contributing to gastrointestinal health[8]. AI is also an emerging transformative force used across various fields, including potential improvements in healthcare diagnosis and treatment associated with large language models (LLMs). There is a specific diagnostic tool, DxGPT, based on LLMs powered by generative pretrained transformer (GPT)-4. By integrating extensive medical knowledge and the language processing capabilities of modern LLMs, DxGPT aims to support clinicians in making accurate and timely diagnoses[9]. Meanwhile, AI is capable of processing large datasets, providing valuable insights into the patterns of emotional experiences and their psychological consequences[10,11].

Language is a unique human endowment that can either amplify or mitigate emotions, leading to deterioration or improvement in mental health[12]. Emotional states and their intensity can be quantified by psychometric tools, which can help evaluate the correlation between IBS symptoms and specific emotions[13]. Finally, psycholinguistic techniques, which analyze the language people use to express emotions, can provide a deeper understanding of how emotional expression may be linked to gastrointestinal health outcomes[14].

Despite growing recognition of this mechanism, understanding and diagnosing the substantial connections between emotional impacts and gastrointestinal health are challenging due to the complex and multidimensional nature of emotional upheavals and their manifestations in the human body[15]. Therefore, this article aimed to explore these innovative methodologies, namely AI, psychometrics, and psycholinguistics, to investigate the influence of emotions on gastrointestinal health. We aimed to minimize the vacuum between subjective emotional experiences, their objectives, and the psychological impact that causes IBS. This approach not only enhances the accuracy of diagnosis but also offers the potential for more personalized and effective interventions, ultimately contributing to better management of gastrointestinal disorders influenced by emotional upheavals.

Significance of the study

The fundamental novelty of the study lies in the interdisciplinary integration of AI DxGPT, psychometrics, and psycholinguistics to examine emotional expression and gastrointestinal health that has not been previously diagnosed[16,17]. The application of DxGPT is unique in analyzing emotional stance, linguistic expressions, and medical records for gastrointestinal health[18]. Language attitude is also rarely used as a predictor of health outcomes, highlighting the psychosomatic consequences of language[19]. The study explains that quality of life mediates emotional expression and gastrointestinal health while coping strategies and resilience moderate the negative consequences. This complex model of independent variable, mediator, dependent variable, and moderators has been hardly tested in psychology so far but offers a new dimension towards conceptual clarity on how emotions deteriorate physical health[20,21].

Emotional narratives (psycholinguistic techniques) guided by the gut-brain axis theory apprehend cultural, emotional, and linguistic nuances, highlighting that local cultural and linguistic practices influence gut-brain dynamics, making results locally relevant and globally transferable[22,23]. Emotional stance, medical records, and linguistic patterns were analyzed using DxGPT to anticipate gastrointestinal deterioration, which is a new contribution to a data-driven, emotionally aware diagnostic paradigm. Further, this study reinforced existing evidence on emotion-oriented gut health, representing a unique methodological synthesis and AI diagnostic intervention to fundamentally contribute to psychological and computational health science. A new preventive diagnostic pathway was proposed, utilizing AI, language markers, and psychometric scales to facilitate early predictions of gastrointestinal issues. This study is both theoretically and clinically cost-effective and translational within the healthcare system.

Hypothetical presentation of the study

Language attitude has a positive association with quality of life (hypothesis 1): Bourhis and Sachdev[24] analyzed vitality perceptions and language attitudes and concluded from a subjective vitality questionnaire and sociolinguistics survey that this combination proved to be a powerful instrument for studying the dynamics of ethnic group perceptions and language attitudes in multilingual settings. Noels et al[25] studied ethnolinguistic vitality and language variation, and the writers suggested three sets of indices. One of them was attitude, and it was included minimally to measure the ethnolinguistic orientation of the speaker. Dragojevic[26] studied language attitude and confirmed that it was organized along two evaluative dimensions: Solidarity and status. He concluded that language attitude was socialized by media, educators, peers, and family. Li and Wei[27] studied the measurement of language attitude constructs and their associations with language achievements. They concluded that the predictive effects of language attitudes on actual language achievement and self-perceived language proficiency were largely confirmed. They finally highlighted the need to incorporate it as a significant individual construct in language learning.

Positive emotional expression is positively associated with quality of life (hypothesis 1a): Michael et al[28] studied the expression and ambivalence regarding the expression of negative emotions in cross-sectional associations with psychological factors and health-related quality of life in postmenopausal females, concluding that ambivalence about expressing negative emotions was more highly correlated with psychological factors and health-related quality of life than emotional expression. They also explained that their results supported the suggestions of prior studies regarding health-related outcomes rather than expressions. Kato et al[29] analyzed positive attitude, emotional expression, self-related health, and depressive symptoms in centenarians and near-centenarians, and path analysis supported the partially mediating role of positive attitude toward life in the relationship between self-related health and depressive symptoms. van Kleef et al[30] suggested that expressions of happiness in coaches were conducive to team performance, and their results highlighted evidence for the benefits of coaches’ positive emotional expression in supporting team performance.

Negative emotional expression is negatively linked with quality of life (hypothesis 1b): The study by Porter et al[31] observed the ambivalence over emotional expression in patients with gastrointestinal cancer and their caregivers. They indicated emotional regulation as a significant factor in understanding patients’ experience and coping with pain. Kuppens et al[32] examined the frequency of positive and negative emotions in relation to life satisfaction across nations and concluded that the emotional aspects of a good life vary across national cultures. Bryan et al[33] investigated the religious coping moderation between association and ambivalence over emotional expression and depression and concluded that there is a need for further examination of ambivalence over emotional expressions. Geng et al[34] examined the relationship between negative emotions and quality of life among adolescents, and the results revealed that negative emotions were negatively associated with quality of life. Rizzo et al[35] observed the link between stigmatization, mental health, disability, and quality of life and explained that quality of life and social integration can be enhanced for individuals with mental health disabilities.

Quality of life positively mediates the relationship between language attitudes and positive emotional expressions and negatively mediates the relationship between negative emotional expressions and general health (hypothesis 2): The study by Stanton et al[36] concluded that emotionally expressive coping predicts psychological and physical adjustment to breast cancer, and the analysis suggests that dispositional hope may be a significant factor in expressive coping as a successful vehicle for goal pursuit. Hevey and Wilczkiewicz[37] studied the language change that mediates the benefit of expressive writing on health-related quality of life following myocardial infarction. Their findings suggest that a higher range of expressive writing is associated with improved health-related quality of life, and expressive writing is also a beneficial intervention for enhancing health-related quality of life among cardiac patients. Kato et al[29] evaluate positive attitudes toward life, emotional expression, self-related health, and depressive symptoms, and the findings indicate the significant role of positive attitudes and emotions, along with self-relation, in mental health. Maalouf et al[38] analyze the personality traits and quality of life among Lebanese medical students, mediating the effects of emotional intelligence, and suggest that the research sheds considerable light on the nexus of associations between emotional intelligence, personality traits, and wellbeing. Nonetheless, it raises more puzzling questions. Overall, it appears that emotional intelligence and its components can serve as an evaluation instrument for assessing the personality profile of Lebanese medical students, providing a basis for future training to enhance the quality of life during medical education.

Resilience moderates positively between negative emotion expressions and quality of life with resilience having a greater impact on the expression of negative emotions (hypothesis 3): Eldeleklioğlu and Yildiz[39] evaluated the expression of emotions, resilience, and subjective wellbeing using structural equation modeling and concluded that there were positive consequences. Armstrong et al[40] investigated emotional intelligence and psychological resilience in response to negative life events, suggesting that class membership fluctuates as a function of four emotional intelligence dimensions with higher scores predicting membership in the resilient class. Groarke et al[41] viewed the quality of life and adjustment for males with prostate cancer. They concluded that perceived stress is the most powerful predictor in the model, and the findings indicated that psychological resilience was a predictive factor that buffered the negative emotions associated with stress. Xie et al[42] elucidated the mediating effects of emotional regulation on the relationship between psychological resilience and distress in young and middle-aged patients with lymphoma, finding that difficulties in emotional regulation partially mediated this relationship. They further explained that enhancing resilience and improving emotional regulation can help alleviate distress by emphasizing the need for targeted psychological interventions in these patients.

Negative emotions deteriorate IBS, and positive emotions soothe IBS by contributing to resilience (hypothesis 4): Parekh et al[43] claimed that IBS is a nonspecific inflammatory chronic disease that is etiologically ambiguous, associated with ulcerative colitis, and has been increasing in numbers. The use of natural language processing elements for computer-aided diagnostic and monitoring of body image perception in orally fed patients with head and neck cancer was evaluated by Różańska et al[44]. The conclusion demonstrated the potential utility of integrating natural language processing fundamentals in patients with psycho-oncological conditions or upper gastrointestinal tract cancers. This method provides a novel approach to assessing the emotional state of the patients. Bao et al[45] analyzed the factors related to spiritual psychology and quality of life in patients with inflammatory bowel disease using AI. They found that patients with inflammatory bowel disease are prone to a combination of obsessive-compulsive, interpersonal, depressive, hostile and other abnormal psychological symptoms. Further, they indicated that their quality of life was significantly reduced and may be affected by the disease condition. Paranoia, depression, and interpersonal issues led to a low quality of life. Madva et al[46] explained the positive psychological wellbeing and clinical features in IBS and concluded that positive psychological wellbeing was diminished in individuals with IBS compared with the general population. They also indicated that greater positive psychological wellbeing was associated with superior physical and psychological outcomes[47].

Coping strategies strengthen the relationship by their moderating effect on the relationship between negative emotions and quality of life (hypothesis 5): Barberis et al[48] investigated the relationships between coping, emotional regulation, and quality of life in patients undergoing dialysis, confirming the link between emotional regulation, coping, and quality of life. The results highlighted the need for comprehensive care of patients, including an assessment of both physical state and psychological functioning, to promote overall wellbeing. Geng et al[34] explored the emotions and quality of life in adolescents using moderating and mediation models in which individual quality of life was indirectly affected by negative emotions via resilience and less affected by negative emotions but more affected by resilience with the improvement of social support. Awad et al[49] assessed the psychometric properties of coping strategies inventory short forms and examined the relationships between resilience and coping strategies, considering the moderating role of emotional regulation. The findings were distinct and novel in all their parameters regarding the relationships between resilience, coping strategies, and emotional regulation. It also brought to light how a population reacts and receives its reality under extreme stress[50].

MATERIALS AND METHODS

Method and participants

The study employed a cross-sectional observational mixed-methods inquiry embedded in routine psychology and gastroenterology care across 20 urban and peri-urban clinics in Pakistan where gastroenterologists were treating the disease. In the sample there were certain variations in the number of participants, ranging from 6 to 17. It explored psychometric assessment, language attitude, and AI-driven DxGPT analysis. A total of 250 adults were enrolled between the ages of 18-40 (average: 29 years) via a purposive sampling technique to ensure the sufficiency of data on patients with IBS and gastrointestinal-related complaints. A battery of validated psychometric instruments was completed by the participants, including language attitudes, positive and negative emotional expressions, resilience, coping strategies, and gastrointestinal-specific quality of life (IBS Quality of Life). We also invited participants to compose a short narrative (150-250 words) explaining their emotional state and current symptoms in relation to these measures.

A subset of 93 participants also consented to link their de-identified clinical notes and medical records, creating a corpus of medical text to analyze through AI. We removed personal identifiers from the clinical data of the patients using a principles-based de-identification workflow supplemented by regular expressions. We also conducted a random audit of human privacy checks to ensure that all personal details (name, address, contacts, age, and dates) had been fully excised. A code book was prepared to ensure annotation for the guidance of human and machine interpretation that consisted of intensity and emotional categories such as rumination, sadness, anxiety, anger, and hopelessness, linguistic markers known as hedging, negation, first person pronouns, temporal framing, and absolute terms, gastrointestinal symptoms termed bloating, irregular bowel habit, nausea, fatigue, and abdominal pain, and cues of resilience and coping called social support problem solving and reappraisal. A clinician adjudicated two trained coders and disagreements in the independently labeled sample of text. Before proceeding to full coding, we ensured the acceptable inter-rater reliability through Cohen’s kappa.

Setting of the study

The study was conducted in Pakistan. We achieved the approval for the study from the Advanced Study and Research Board of The Islamia University of Bahawalpur (Pakistan) on September 22, 2024. We collected and analyzed the data from October 2024 to May 2025. Finally, the study was completed in June 2025. All the activities were ethically standardized throughout the whole time period.

Instruments

The study used various scales to measure emotional expression and the IBS Quality of Life scale for gastrointestinal health. Initially, we administered a structured questionnaire to collect fundamental data on various demographic factors, such as age, socioeconomic status, marital status, personal history of illness, length of treatment, brief family and personal history, duration of disease, time of treatment, and improvement. Furthermore, the data were collected using a questionnaire with different Likert scales. We used the positive and negative emotional expression scales separately[51]. Both emotional expressive scales have ten items and four factors, and they were measured on five-point Likert scales. Then, we used the language attitude scale, which has ten items and three factors, and was measured on a seven-point Likert scale[27]. After that we measured quality of life using the scale from Andrae et al[52]. It has ten items, one factor, and yes/no questions. Elklit’s scale[53] was also used to measure coping strategies. It has 20 items, one factor, and a four-point Likert scale. Finally, the van der Meer et al[54] resilience scale was used, consisting of ten items, two factors, and a five-point Likert scale. The qualitative data on emotional expression were studied through content analysis guided by the gut-brain axis and relevance theories[55,56].

Qualitative procedure

The AI pipeline connected an LLM with the layer of classical machine learning. The GPT-4 class model was prompted in the LLM layer with a codebook and a handful of exemplars for each category. A constrained Java Script Object Notation schema was used to extract structured outputs from free text, and it included emotion categories intensity, symptom mentions normalized to a gastrointestinal lexicon, the presence of coping resilience, and language attitude. Simultaneously, we also engineered psycholinguistic features, including token counts, pronoun ratios, type-token ratios, model verbs, question frequencies, negations, stance markers, hedges, temporal framing, and linguistic inquiry word count-style affect dictionaries. We also used trained gradient and logistic boosting models for the prediction of high against low IBS Quality of Life impairment, depression, and anxiety flags and symptom clusters. We adopted soft voting and probability calibration for the outputs from the two layers[57,58]. Performance was calculated through five-fold cross-validation on the 93-record subset. A panel of two gastroenterologists and one health psychologist were consulted to ensure the authenticity of the AI conclusions. They reviewed the AI summaries of sampling for usefulness and plausibility. DxGPT AI also analyzed the medical records and emotional stances of participants. Finally, qualitative content analysis was conducted guided by the gut-brain axis and relevance theories, utilizing code-recode reliability checks and negative case analysis to challenge early themes.

Quantitative procedure

Quantitative data were analyzed using SmartPLS software. We proceeded with quantitative modeling by the validity and reliability check of all latent constructs (Cronbach’s alpha, average variance extracted, composite reliability, and heterotrait-monotrait for discriminant validity), followed by partial least squares-structural equation modeling bootstrap resample to measure mediation (quality of life as the conduit between gastrointestinal outcomes, emotions, and language attitude) and moderation (coping and resilience as amplifiers). The standard covariates, including gender, age, socioeconomic status, marital status, and residence (urban/rural), were also included. They were further inspected for multicollinearity through the variance inflation factor. Furthermore, we analyzed common method variance using Herman’s single-factor test and full collinearity variance inflation factor as self-report can inflate correlations[59]. We simultaneously conducted a sensitivity analysis stratified by language response (Urdu vs English) and by clinic stratum (urban vs peri-urban). Finally, we triangulated the results with psychometric patterns indicated through SmartPLS, utilizing AI-derived features.

Data structure limitations and sampling bias

We adopted an explicit strategy to address biases introduced in our recruitment, sampling, data collection, and analysis. The purposive clinical-based sampling privilege may help individuals who are already under treatment, potentially underrepresenting those with limited access, different help-seeking norms, and subclinical symptoms. Our sample was more urban and more educated, and the subset of 93 individuals’ AI records was also constrained by data availability. It enhanced the possibility of minimizing the chance of missing at random selection in AI analysis. However, we attempted to mitigate these issues by reporting stratum-specific estimates (urban vs rural), adjusting socioeconomic covariates in the structural model, and comparing non-responders and responders through key demographic and IBS quality of life variables. We also used conservative imputation strategies when it was appropriate with AI error profiles by age, gender, and urbanicity by identifying systematic misclassification. Finally, we documented the thresholds, prompts, and model version for enabling replication.

Ethical concerns

We obtained informed consent in Urdu and English languages; data were collected through anonymous online surveys that only proceeded after consent with the right to withdraw from the research without any consequence. We separated the research process from clinical care to avoid therapeutic misconception. The anonymity of participants was ensured, and the medical record was scrubbed of identifiers before analysis. The AI process had a chance of human oversight. Therefore, AI results were considered an interpretive strategy. The study was conducted under the Advanced Study and Research Board of The Islamia University of Bahawalpur (Pakistan) approval with a data-management and AI/privacy impact plan, preregistered analyses, disclosure of conflicts of interest, and a commitment to aggregate and responsible dissemination.

Statistical analysis

We employed a comprehensive statistical data analysis procedure to ensure the validity and robustness of the findings. SPSS version 31 was used to generate descriptive statistics on the participants’ demographic traits. The model evaluation and hypothesis testing were conducted using SmartPLS 4 because it was well-suited for analyzing complex models with moderating effects and multiple latent constructs. A two-step approach was followed in the analysis, beginning with the assessment of the measurement model, which evaluated reliability, convergent validity (Cronbach’s alpha), internal consistency, average variance extracted, and composite reliability. After confirming that the acceptable thresholds were met (α > 0.70, composite reliability > 0.70, average variance extracted > 0.50), we tested the structural equation modeling (SEM) to examine the hypothesized relationships among constructs. The direct and moderating effects were assessed using SEM with partial least squares, and bootstrapping with 5000 resamples was used to determine the statistical significance of path coefficients, t values, and P values, providing robust estimates for hypothesis testing. Moreover, the conditional effects analysis was performed to evaluate moderating effects by examining the relationships between the dependent and independent variables. Finally, the precision and reliability of the moderation results were assessed using confidence intervals (set of 93%).

RESULTS

Quantitative findings

Initially, we used SPSS version 31 for the statistical analysis of the demographic data; however, SmartPLS (https://www.smartpls.com) was also employed[60]. This software was used for all other analyses. Then, we used a structural equation model to analyze the moderation of the dependent variable, which was gastrointestinal health, and to examine the roles of coping strategies and resilience in the relationships between quality of life, language attitude, and positive and negative emotions. After that we initiated conditional effects of testing, such as the influence of one variable over another when conditioning on a third variable and other effects. It also measured the effect of X on Y at different points along with the moderator by analyzing whether this effect was significant. Finally, we established the statistical importance of simple moderations at 93% confidence intervals (Figure 1).

Figure 1  Path diagram.

Demographic

We took 250 participants from different clinics including 107 males and 143 females with an average age of 29 years. They belonged to middle class families. Most of them were educated but unemployed. Demographic details can be viewed in the Table 1.

Table 1 Demographics.

Factors	Number (n = 250)	Frequency, %
Age
18-23	65	26.0
23-27	83	33.2
28-35	58	23.2
35-40	44	17.6
Gender
Female	143	57.2
Male	107	42.8
Study level
Bachelors	157	62.8
Masters	93	37.2
Residency
Peri urban	93	37.2
Urban	157	62.8
Total	250	100
Marital status
Married	99	39.6
Unmarried	143	57.2
Unknown	8	3.2
Total	250	100

We assumed that Cronbach’s alpha, composite reliability, and average variance extracted were necessary metrics for assessing the internal consistency, reliability, and convergent validity of the latent construct. By using these metrics the researchers ensured the robustness of the measurement, thereby enhancing the accuracy of the overall analysis. All three are visible in Table 2. We calculated the Cronbach’s alpha value for each construct to measure reliability, and it ranged from 0.75 to 0.94 across all constructs, exceeding the minimum threshold of 0.7. Then, we analyzed the reliability of each variable using composite reliability, and it resulted between 0.84 and 0.95, constantly crossing the threshold of 0.7. Finally, average variance extracted was calculated to confirm convergent validity, and it exceeded the threshold value of 0.5.

Table 2 Cronbach’s alpha, composite reliability, and average variance extracted for each latent construct.

Number	Variable name	Cronbach’s alpha	CR	AVE
1	Coping strategies	0.794	0.902	0.592
2	Generalized health deterioration	0.771	0.882	0.557
3	Language attitude	0.872	0.928	0.604
4	Negative emotional expression	0.864	0.845	0.534
5	Positive emotional expression	0.880	0.898	0.665
6	Quality of life	0.948	0.846	0.576
7	Resilience	0.750	0.952	0.627

CR: Composite reliability; AVE: Average variance extracted.

The strength and direction of relationships between variables were presented through a standardized effect table in homogeneous manners, providing a clear reference for assessing their relative influence in a statistical model. Table 3 depicts the contribution and significance of various variables in the analysis. The path and effects of a significant negative relationship are depicted in Table 3, showing the relationship between negative emotional expression and quality of life, resilience, and coping strategies as well as negative emotional expressions. Then, the moderator’s coping strategies had a positive influence on the mediator’s quality of life. Independent variables, such as language attitude and positive emotional expressions, also had a positive influence on the mediator’s quality of life. The quality of life mediated positive influence on the dependent variable (gastrointestinal health).

Table 3 Path coefficients, t-values, and P-values for variable relationships in the statistical model.

Number	Variable name	Path coefficients	t values	P values
1	Coping strategies affects quality of life	0.044	3.855	0.000
3	Language attitude affects quality of life	0.693	9.161	0.000
4	Negative emotional expression affects quality of life	-0.112	2.191	0.029
5	Positive emotional expression affects quality of life	0.243	3.258	0.001
6	Quality of life affects gastrointestinal health	0.035	14.526	0.000
7	Resilience affects quality of life	-0.144	7.466	0.000
8	Resilience × negative emotional expression	0.062	7.521	0.000
9	Coping strategies × negative emotional expression	-0.014	12.450	0.000

Then, we analyzed the direct and indirect path estimates as well as the moderating effects of the study model. We ran the structural model through bootstrapping (n = 5000) to generate path coefficients (β), t values, and P values. The t results provided evidence supporting most of the hypotheses. Language attitude showed a strong positive association with quality of life (β = 0.693; t = 9.161; P < 0.000) while positive emotional expression also had a significant positive effect on quality of life (β = 0.243; t = 3.258; P = 0.001). In contrast, negative emotional expression had a significant negative relationship on quality of life (β = -0.112; t = 2.191; P = 0.029). Furthermore, coping strategies were positively linked to quality of life (β = 0.044; t = 3.855; P < 0.000). Quality of life significantly predicted gastrointestinal health (β = 0.035; t = 14.526; P < 0.000), suggesting a mediating path. Resilience showed a negative direct effect on quality of life (β = -0.144; t = 7.466; P < 0.000), but its interaction with negative emotional expression was significantly positive (β = 0.062; t = 7.521; P < 0.000), confirming a moderating role. The interaction between coping strategies and negative emotional expression also showed a significant negative moderating effect (β = -0.014; t = 12.450; P < 0.000).

Qualitative findings

Then, we analyzed the medical stance of patients’ clinical observations and medical records using a DxGPT-style analysis of 93 patients as the rest of the sample either lacked records or had lost them due to certain reasons. We found from the record that most patients were associated with disordered eating patterns, anxiety, negative emotional stance, and frustration. Of the 93 patients 87 were diagnosed with IBS, and 6 patients were suffering from Helicobacter pylori infection. The patients with IBS were suffering from chronic abdominal discomfort, bloating, irregular bowel habits associated with low appetite, loss of sleep, low mood, and frustration caused by multidimensional circumstances. The percentage of common comorbidities causing IBS found on DxGPT were generalized anxiety disorder (29%), somatic symptom disorder (6%), depressive disorder (26%), sleep disturbance (insomnia) (14%), and eating disorder patterns (15%). We analyzed the symptoms clustering record of patients with IBS from their documents, which indicated abdominal discomfort in 25%, irregular bowel habits in 32%, bloating in 9%, fatigue in 18%, and nausea in 16%. Finally, the results supported the hypothesis that emotions predict gastrointestinal deterioration, and it was consistent with the associations between gastrointestinal outcomes and emotional measures that causally cannot be inferred[61].

DISCUSSION

The results of the study suggest that language attitude plays a significant role in shaping individuals’ perceptions of their quality of life. Individuals experience a higher sense of belonging and empowerment when they view language positively, particularly in contexts when language proficiency contributes to achievement, identity, and social inclusion. It is parallel to forming a language attitude for sociocultural adoption as noted by Yi and Zhang[62], who confirmed that language attitude presents the fundamentals of sociocultural adoption. This psychological empowerment likely enhances emotional wellbeing and life satisfaction. The findings illuminate acculturation frameworks and human capital in which language serves as either a means of communication or symbolizes self-worth, competence, and integration in extensive sociocultural settings. It is aligned with the study by Kartini et al[63] that concluded that integration of local cultural elements enhances a greater connection to language learning content.

The quality of life is influenced by the positive emotional expression of individuals, serving as another key predictor. Positive emotions are frequently expressed, supporting a higher standard of life satisfaction, aligning with psychological theory, and associating with optimism, improved interpersonal relations, and psychological resilience. This finding is consistent with the results of Wang et al[64], who suggest that positive emotions have a direct and significant effect on life satisfaction. Positive emotions are a strategy that enhances the capacity to think constructively by maintaining social bonds and facing daily life challenges with flexible perceptions. However, negative emotion expression is associated with handicapped quality of life, and the findings are consistent with the effective processing model. It indicates that chronic negative expressions disrupt social harmony, drain emotional energy, and accelerate cognitive bias towards distress and dissatisfaction, which can affect gastrointestinal health.

The study further exposed that quality of life was a significant intermediary between gastrointestinal wellbeing and emotional experiences. It is parallel to the study by Han et al[65] that demonstrated that the mechanism of mental health affected the occurrence and development of gastrointestinal cancer. Individuals who possess a favorable language attitude and a positive emotional stance tended to experience a better quality of life and less deterioration of their gastrointestinal health (IBS). This finding supports the biopsychosocial view of health in which subjective wellbeing serves as a conduit through which psychological and emotional processes influence physical conditions[66]. It also supports the theory of the gut-brain axis by recognizing contemporary health psychology, which explains how emotional mechanisms influence the function of the human body, including internal stress responses and digestion[67].

The study highlighted resilience as a crucial moderator that impacts negative emotional expression on quality of life. Individuals who exhibited higher resilience tended to experience fewer declines in life satisfaction even when expressing negative emotions. It supports the resilience theory, which posits that internal strengths and adaptive coping mechanisms enable individuals to rebound from adversity, maintain equilibrium, and protect their wellbeing in the face of emotional turbulence[68]. Instead of being overwhelmed by the negative consequences, resilient individuals may compartmentalize and tolerate the stress to avoid long-term psychological issues.

Coping strategies play a crucial moderating role; however, their influence is complex and multifaceted. The interaction between coping and negative emotional expression revealed that all coping is not equally effective. It depends on the management of individuals’ emotional stress, whether it is problem-solving, rumination, or a combination of both. Furthermore, the coping strategies may either exacerbate or minimize the consequences of negative emotions on quality of life. It reflects the transactional model of stress and coping[69,70], which explains that the outcome of coping is contingent upon its alignment with the nature of the stressor. The emotionally amplified strategies and unintentional emotional burden can alleviate the stress.

Our results align with the modern reframing of IBS and associated disorders of gut-brain interaction, emphasizing bidirectional signals between the enteric and central nervous systems[12]. The study built upon the work of Nisticò et al[71], who claim that neuroimaging findings revealed alterations in the insula and anterior cingulate in IBS, mapping interoceptive and affective networks to the severity of symptoms.

The study also supported the findings that therapeutic interventions (psychotherapies and medicine) targeting the gut-brain link are an appropriate strategy for better health among affected patients[72]; however, there may be variations in the results from person to person. The results aligned with those of Osadchiy et al[73] and Mayer et al[74], who indicated that bidirectional gut-brain interactions, including immune regulation, food intake, and sleep regulate key homeostatic and physiological functions. The results further indicated that stress and the immune system interact in light of psychoneuroimmunology, which explains how emotions affect gut immunity and sensitivity. Therefore, our findings are consistent with the hypothesis linking emotional expression, quality of life, and gastrointestinal health[75]. Finally, our study may enhance clinical management. Further, perspective studies will improve the test reliability, subgroup fairness, and impact before implementation.

Limitations of the study

We consider this clinic-based purposive sample as a single time point “snapshot” in the cross-sectional study because residual confounding, such as diet, Helicobacter pylori results, medications, socioeconomic status of stressors, comorbid anxiety, reverse causation of gastrointestinal symptoms, and negative effects can be reflected through associations; therefore, causal inference and temporal ordering cannot be established. Further, there is also a common method variance of mood effects and recall bias. Furthermore, cultural linguistic adaptation may concern measurement invariance, and a psycholinguistic sample can encode a dialectical change that shifts meanings. Finally, the results are framed as associative, and in future studies longitudinal research design, preregistered sensitivity analysis, clinical anchors, and error audits are strongly recommended because mediation is interpreted as a subsidiary suggestion and moderation is exploratory. Therefore, residual confounding and missing not at random risk in the AI subset are acknowledged.

Revolutionary innovation

There is an integrated diagnostic paradigm that highlights the study’s revolutionary innovation by combining AI, psycholinguistics, and psychometrics to identify and predict deterioration in gastrointestinal health through emotional expression. Contrary to routine diagnostic interventions, which rely solely on psychological markers, the study represented a revolution in adopting an emotion-centered approach and AI-driven tools like DxGPT to analyze integrated emotional, linguistic, and psychological data alongside medical records. The research established a gut-brain emotional intelligence framework by correlating psychometric scales, AI-detected emotional patterns with gastrointestinal outcomes (the IBS Quality of Life scale), and language attitude. This interdisciplinary fusion enables early, noninvasive detection of emotional precursors to gastrointestinal disorders, offering a transformative pathway for precision psychosomatic medicine and emotional health-driven gastroenterology.

CONCLUSION

The study highlighted an intricate interplay between language attitude, emotional stance, resilience, and coping mechanisms, which determine psychological and physical wellbeing. The results were found through the examination of different pathways that contribute to comprehension of the psychological process of health deterioration caused by language and emotions. The study will be helpful to reduce emotion-induced gastrointestinal problems. Our results motivate the validation of future research. We do not recommend any clinical implementation until further assessment in which emotional issues co-occur with gastrointestinal health.