Online self-disclosure discourse among individuals with depression in China: an analysis of “depression super topics” on Microblog

3.3.1 Data preprocessing

Initially, we import and process the collected corpus using “Pandas” in Python. To ensure the accuracy of the data, basic cleaning tasks are conducted, including the removal of unnecessary symbols, spaces, and other special characters that might affect the analysis. Given that the comments are expressed in Chinese, we utilize “Jieba” (Chinese text segmentation: built to be the best Python Chinese word segmentation module) to remove irrelevant punctuation and whitespace information. This step is crucial for accurately conducting the corpus analysis in this study.

3.3.2 Word frequency analysis

Word frequency analysis is a content analysis method used to identify the focus and key words in a text (Mao et al. 2023). First, in terms of linguistic features, this study primarily employs word frequency analysis. Specifically, regarding the usage patterns of personal pronouns, we use Python’s Pandas library in combination with Excel tools to conduct the analysis and tabulation. Initially, we load the data into a DataFrame using “read_excel” function from the provided Excel file, facilitating subsequent data processing and analysis. Next, we define a list of eight Chinese personal pronouns (“I”, “We”, “You”, “He”, “She”, “It”, “They”) and tally the occurrence of each pronoun in each statement within the “Statement Content” column of the DataFrame. enabling efficient data processing. To examine lexical usage, we also conducted word frequency analysis on the study corpus, with the results visualized as a word cloud generated in Python.

3.3.3 Sentiment analysis

Sentiment analysis involves the extraction of subjective information from text using techniques such as text mining. Currently, there are two primary methods for conducting sentiment analysis: one is based on sentiment lexicon, and the other employs machine learning techniques (Peng et al. 2024). The lexicon-based sentiment classification approach utilizes well-established sentiment dictionaries to calculate the emotional value of sentiment words found in the text, which then determines the overall sentiment polarity of the text through weighted calculations (Wang et al. 2022). This study employs the Affective Lexicon Ontology (Xu et al. 2008) developed by Dalian University of Technology. However, recognizing the limitations of this approach, which include an over-reliance on the construction of sentiment dictionaries and the inability to consider context and situational nuances, this research aims to optimize by integrating Python and Gephi software to visualize the co-occurrence network of prominent sentiment word frequencies, thereby further verifying the accuracy of the results.

3.3.4 Topic analysis: based on LDA model

Topic modeling, particularly the Latent Dirichlet Allocation (LDA) model, is widely used in text data mining to uncover latent semantic relationships between documents and words (Ding and Wu 2023). The LDA model treats documents as a mixture of topics, where each topic is a probability distribution of words, enabling efficient text information processing. In this study, the LDA model is employed to analyze the implicit semantic structures of online self-disclosure discourse among depression patients on microblog. The model identifies the distribution probabilities of topics within documents and terms within topics (Fan and Liu 2023), grouping related terms to represent corresponding themes (Chen et al. 2022).

The LDA model requires the determination of the optimal number of topics, K, to ensure effective topic extraction (Ding and Wu 2023). In this study, perplexity is calculated to determine the optimal number of topics, with lower perplexity indicating better clustering performance. As shown in Figure 1, when the number of topics is set to 5, the perplexity of the topics mined by the LDA reaches a local minimum, indicating optimal clustering results at this topic number.

Figure 1:

Confusion degree change of topics.

4.1.1 Pronoun usage characteristics

This section aims to conduct a detailed analysis of personal pronouns used in the online discourse of individuals with depression on social media. Specifically, it aims to analyze the frequency of first-person, second-person, and third-person pronouns (see Table 1) in the corpus to examine the degree of self-focus (see Figure 2) in the context of Chinese online self-disclosure among individuals with depression. By exploring the patterns of personal pronoun usage in these self-disclosures, this study seeks to uncover how such linguistic choices reflect the psychological states of individuals with depression and shape their interpersonal interactions in online communication.

Figure 2:

Online self-disclosure discourses related to “我” (wo, ‘I’) by individuals with depression.

The results from Table 1 demonstrate the usage patterns of personal pronouns in the online self-disclosure of individuals with depression. First, the most frequently used pronouns are first-person pronouns, with a frequency of 1242, among which the first-person singular “I” registers the highest frequency across all categories (Huang and Zhou 2021); this is followed by third-person pronouns with a frequency of 320, and second-person pronouns with a frequency of 215. This indicates that the online self-disclosure discourse of those with depression in China on social media also tends to be highly self-focused (Ding and Wu 2023), which will be analyzed in detail in the following sections (see Figure 2). Notably, as an affective positioning resource (Giaxoglou 2015), the use of third-person pronouns in Chinese online depression discourse constructs a state of isolation in terms of external identity and internal emotions. Furthermore, third-person pronouns serve to reinforce a confrontational context, where individuals describe their emotional experiences from an external perspective while simultaneously linking negative evaluations to the discourse of others, thereby creating a tension and conflict between self-position and external discursive frameworks.

In Example (1), the speaker uses the third-person plural “他们” (tamen, ‘they’) to describe their relationship and communication with their parents. This usage linguistically constructs a clear identity opposition and reflects significant social and emotional distance between the speaker and their parents. Initially, by categorizing the parents as “他们” (tamen, ‘they’) the speaker not only emphasizes a physical separation but also expresses an opposition and estrangement on a psychological and emotional level (Hu 2023). Moreover, attributing the emotional conflicts between “我” (wo, ‘I’) and parents to an external“他们” (tamen, ‘they’) is a strategy of externalization employed by the speaker, demonstrating a mechanism of self-protection. The interaction described by the speaker is filled with accusations and criticisms, such as the parents unfavorably comparing the patient to others and labeling them as “敏感” (mingan, ‘sensitive’). In describing the actions and words of the parents, the speaker uses the third person to maintain a certain degree of objectivity and distance (Liu and Li 2012), thus helping the patient to emotionally delineate boundaries. Through this approach, the speaker not only protects themselves from direct harm but also attempts to voice their complaints within the online community (Zhang and Shuai 2023), seeking empathy and support.

Second, the use of singular pronouns (frequency: 1688) significantly surpasses that of plural pronouns (frequency: 89) in the online self-disclosure language of individuals with depression. This indicates that self-disclosure on social media by those with depression is often more personalized (Sun 2023). The frequent use of singular pronouns may be related to their expression of personal emotions and requests for help, while the less frequent use of plural pronouns (e.g. “我们”, ‘we’ frequency: 35) suggests that building interactions and resonances with others on social media is relatively limited. Specifically, the first-person singular “我” (‘I’) (frequency: 1207) is the most frequently used, which re-emphasizes the self-focused nature of the language used by those with depression in articulating their feelings and experiences (Ding and Wu 2023) and the introspective monologue style of their speech.

In Example (2), the individual with depression emphasizes his or her personal experiences and emotional state by consistently using the first-person singular “我” (‘I’), as seen in statements such as “我生病了” (wo shengbing le, ‘I’m sick’), “我得了抑郁症” (wo dele yiyuzhheng, ‘I have been diagnosed with depression’) and “我好痛苦” (wo hao tongku, ‘I am in so much pain’). Such expression not only demonstrates how the individual connects life’s setbacks and personal illness to form a negative life narrative, but it also reflects the personalized and introspective monologue style of language used in self-disclosures on social media. This is a primary linguistic characteristic of how this group presents themselves online.

To further explore the self-focused content in the Chinese online self-disclosure of individuals with depression, this study conducted a semantic network analysis on the corpus of utterances related to the pronoun “我” (wo, ‘I’). We selected “我” (wo, ‘I’) as the keyword, filtering out 303 relevant comments, and retained texts longer than two words and ranked within the top 70 by significance. The cleaned data was then visualized using Gephi to create a semantic network diagram (see Figure 2). Semantic network analysis allows for the assessment of co-occurrence relationships between words, where the density of the lines indicates the closeness of the relationships, and the denser the lines, the stronger the association between the words (Lin and Meng 2020). As demonstrated in Figure 2, the following three conclusions can be drawn.

First, the online self-disclosure utterances about “我” (wo, ‘I’) among individuals with depression demonstrate their self-focus on emotional states. The central term “我” (wo, ‘I’) is closely linked with various words related to emotions and psychological states, revealing the emotional self-focus in the online discourse of those with depression on social media. Specifically, the links between “我” (‘I’) and terms like “情绪” (xingxu, ‘emotion’), “痛苦” (tongku, ‘pain’), “焦虑” (jiaolv, ‘anxiety’), “害怕” (haipa, ‘afraid’), “难受” (nanshou, ‘discomfort’), and “崩溃” (bengkui, ‘collapse’) highlight the core emotions frequently expressed by those with depression. The dense connectivity of these emotional words illustrates how often individuals discuss their negative feelings on social media platforms (Fan and Liu 2023; Vedula and Parthasarathy 2017).

Second, the co-occurrence network vividly illustrates the focus of individuals with depression on their own life states on social media. The central word “我” (wo, ‘I’) is closely linked with multiple terms related to everyday life, such as “抑郁症” (yiyuzheng, ‘depression’), “世界” (shijie, ‘world’), “活着” (huozhe, ‘living’), “意义” (yiyi, ‘meaning’), “睡觉” (shuijiao, ‘sleeping’), “生病” (shengbing, ‘sickness’), and “状态” (zhuangtai, ‘state’). The frequent appearance of these words reveals the patients’ focus on their own health status, physiological conditions, and daily challenges, and even extends to broader existential questions, such as worldviews and the meaning of life (Ding and Wu 2023).

Third, a range of terms associated with “我” (wo, ‘I’) such as “朋友” (pengyou, ‘friends’), “家人” (jiaren, ‘family’), “老师” (laoshi, ‘teachers’), and “同学” (tongxue, ‘classmates’) prominently reflect the focus of individuals with depression on their interpersonal relationships on social media. Through social media, patients share their personal interpersonal experiences from real life, primarily disclosing challenges they face such as social pressures and feeling marginalized within their social circles (Lian 2020). This also highlights their deep need for understanding and acceptance.

In summary, regarding the use of personal pronouns, the self-disclosure discourse of individuals with depression predominantly employs first-person pronouns, followed by third-person pronouns and second-person pronouns. Pronouns serve as tools for online self-disclosure practices and emotional regulation, achieving effects of emotional immersion and detachment, thus reflecting the strategic use of language by the speaker. Moreover, the discourse of individuals with depression online primarily exhibits a self-focused narrative pattern (Ding and Wu 2023), characterized by a focus on one’s own emotions, life circumstances, and interpersonal relationships.

4.1.2 Lexical usage characteristics

To visually understand the vocabulary features of depression patients’ online discourse on the microblog platform, this study conducted word frequency statistics on the selected corpus and used Wordcloud in Python to generate a high-frequency word cloud (see Figure 3). Figure 3 displays the common vocabulary in the online discourse of the depression community, with larger fonts indicating higher frequency of use of the corresponding words.

Figure 3:

Word cloud of online discourse from the “depression super topics”.

Based on Figure 3, the most frequently occurring words include “感觉” (ganjue, ‘feeling’), “抑郁症” (yiyu, ‘depression’), “情绪” (qingxu, ‘emotion’), “症状” (zhengzhuang, ‘symptoms’), “不想” (buxiang, ‘don’t want’), “焦虑” (jiaolv, ‘anxiety’), “喜欢” (xihuan, ‘like’), “患者” (huanzhe, ‘patient’), “痛苦” (tongku, ‘pain’), “抑郁” (yiyu, ‘depressed’), and “难受” (nanshou, ‘discomfort’). Therefore, the linguistic features of the online discourse of individuals with depression on social media can be categorized into three main aspects: i) frequent use of disease-related vocabulary; ii) frequent use of negative expressions; iii) frequent use of emotional vocabulary.

First, individuals with depression frequently use disease-related vocabulary on social media, such as “抑郁症” (yiyuzheng, ‘depression’), “症状” (zhengzhuang, ‘symptoms’), “患者” (huanzhe, ‘patient’) and “抑郁” (yiyu, ‘depressed’). The high frequency of these words indicates their focus and awareness of their own health conditions. Additionally, through these words, patients not only describe their pathological states but also seek relational connections and identity affirmation within the online community (Hu 2023).

Second, the frequent use of negative terms like “不”(‘no’ or ‘not’) is another significant characteristic of the online self-disclosure language of individuals with depression, often linked to their negative thinking patterns and sense of uncertainty (Yan and Liu 2022). The use of these words typically reflects the patients’ uncertainty about the future, dissatisfaction with their current conditions, or an escapist attitude.

Third, another noteworthy linguistic feature is the frequent use of emotional vocabulary. Emotional words such as “情绪” (qingxu, ‘emotion’), “焦虑” (jiaolv, ‘anxiety’), “喜欢” (xihuan, ‘like’), “痛苦” (tongku, ‘pain’) and “难受” (nanshou, ‘discomfort’) reflect the emotional fluctuations and psychological states expressed in the online self-disclosure discourse of individuals with depression (Huang and Zhou 2021).

4.3.1 Topic one: emotional representation

The topic of Emotional Representation primarily addresses how individuals with depression express their emotional fluctuations and psychological experiences online via social media. This topic typically reveals the direct emotional states of the patients and how these emotions impact their daily lives. High-frequency words in this topic, such as “情绪” (qingxu, ‘emotion’), “喜欢” (xihuan, ‘like’) and “难过” (nanguo, ‘sad’) indicate that patients express a wide range of emotions on social media. This not only reflects the volatility of their emotions but also demonstrates the varying intensity of these emotions. For instance, “喜欢” (xihuan, ‘like’) might refer to moments of positive emotion, whereas “难过” (nanguo, ‘sad’) and “害怕” (haipa, ‘afraid’) reflect deeper negative emotional experiences (Herrman et al. 2019). Additionally, words like “裂开” (liekai, ‘break down’), “导致” (daozhi, ‘lead to’), ‘失眠” (shimian, ‘insomnia’), and “影响” (yingxiang, ‘influence’) suggest that patients are concerned not only with the emotions themselves but also with the long-term effects of emotional fluctuations on their health (Fan and Liu 2023). In summary, the self-disclosure discourse under this topic primarily reveals how individuals with depression navigate their complex inner experiences through diverse and intense emotional expressions on social media. These emotional expressions are not only direct reflections of their psychological states but also serve as a means for them to release emotions and alleviate the stress brought on by those emotions.

4.3.2 Topic two: emotional challenges

The topic of Emotional Challenges pertains to the specific difficulties patients encounter when managing their personal emotions and psychological distress. This often includes the struggle against deep emotional pain and the search for ways to cope with such suffering. High-frequency words like “苦涩” (kuse, ‘bitterness’) and “痛苦” (tongku, ‘pain’) highlight the intense inner conflict and struggles faced by patients (Zhang 2022), while “吃药” (chiyao, ‘take medicine’) and “睡觉” (shujiao, ‘sleep’) reflect the practical measures they take to alleviate these emotions, sometimes even extending to extreme solutions like “死” (si, ‘death’). Additionally, high-frequency words like “哭” (ku, ‘cry’) and “不想” (buxiang, ‘don’t want’) may express a psychological state of avoidance and resistance (Zhang 2022), which are natural reactions for patients facing extreme emotional pressure. The online discourse in topic two focuses on how individuals with depression manage intense emotions, revealing the severe health crises they face. Consequently, this category serves as a crucial linguistic marker for mental health professionals when providing support and intervention.

4.3.3 Topic three: life status

The topic of Life Status in the online discourse of individuals with depression focuses on aspects of their daily lives affected by depression, particularly in terms of social interactions, health management, and overall quality of life. Patients often express difficulties related to communication with “朋友” (pengyou, ‘friends’) and “老师” (laoshi, ‘teachers’) on social media, reflecting potential interpersonal relationship barriers they face in social settings (Yan and Liu 2022), as illustrated in Example (4).

Example (4) illustrates the speaker’s discomfort and feelings of alienation during social activities. This social difficulty is not limited to interactions with friends and within educational environments but may extend to a broader social circle, demonstrating how depression can impair patients’ social abilities and participation in society. Additionally, words like “药” (yao, ‘medicine’) and “医院” (yiyuan, ‘hospital’) reveal the reality that patients must rely on the medical system to manage depressive symptoms, while “生病” (shengbing, ‘illness’) further indicates that, beyond mental health issues, patients may also be dealing with other health problems. This suggests that the impact of depression is comprehensive, affecting not only mental health but also overall physical well-being.

4.3.4 Topic four: pathological manifestations

This topic focuses on the physiological impacts and physical symptoms brought about by depression, offering a direct description of the patients’ health status. Depression is not only an abnormal psychological state but also triggers a range of physical symptoms, such as chronic pain, sleep disorders, and digestive issues (Lian 2020). In the health discourse on social media, patients frequently mention issues related to “疼痛” (tengtong, ‘pain’) and “躯体” (quti, ‘somatic’), which not only reflect their daily struggles but also highlight the inseparable connection between mental and physical health, as illustrated in Example (5).

In Example (5), the patient, by describing a series of specific physiological sensations, not only conveys a direct experience of their symptoms but also attempts to seek understanding and potential solutions. In this context, social media serves not only as a channel for expressing and sharing the pain caused by the illness but also as a means to seek concrete solutions.

4.3.5 Topic five: emotional needs

This topic in online discourse highlights the need for emotional support and understanding among depression patients on social media, clearly indicating their intent to seek help and support. High-frequency words such as “爱” (ai, ‘love’), “希望” (xiwang, ‘hope’) and “抱抱” (baobao, ‘hug’) directly reveal their need for emotional comfort (Hu 2023), as illustrated in Example (6).

Example (6) reveals the deep sense of loneliness experienced by the patient, along with a strong desire for emotional support. As a result, individuals with depression use social media platforms to express their emotional needs, attempting to break through the isolation imposed by depression. This further indicates that under the dual pressures of inner turmoil and social barriers caused by depression, the “抑郁症超话” (yiyuzheng chaohua, ‘depression super topics’) online community on social media serves as a crucial channel for social and emotional support. Through this platform, patients can connect with others who may be experiencing similar feelings, which not only helps to alleviate the despair and loneliness felt in real-life social interactions but also contributes to the construction of a supportive online social network (Yan and Liu 2022).

In summary, this section, through the use of the LDA topic model, delves into the online self-disclosure discourse of depression patients on social media, revealing five key topics: emotional representation, emotional challenges, life status, pathological manifestations, and emotional needs. Together, these topics paint a comprehensive picture of how individuals with depression express their complex emotional experiences and life challenges in the digital space, while also highlighting their urgent need for social support and emotional understanding. Notably, beyond the direct negative impacts of their life status and physical health, the emotional representations of fluctuations, crises, and needs are critical factors that cannot be overlooked in the process of supporting and understanding depression individuals with depression.