Linguistic markers of depression and emergent self-stigma in online self-disclosures: A mixed-methods study on...

Depression is a global mental health crisis, exacerbated by stigma that hinders open expression, particularly in collectivistic cultures. While linguistic markers of depression (e.g., self-focus, negative emotions) have been identified in clinical, Western contexts, their replication in non-clinical, everyday online discourse and their intersections with self-stigma remain underexplored.

Using a mixed-methods design, we collected 627 depression disclosures and 729 control posts from the social media platform Xiaohongshu via automated web crawling. Posts were analyzed using the Simplified Chinese Linguistic Inquiry and Word Count (SC-LIWC), a validated text analysis tool, to quantify linguistic features. Following this, a quantitative analysis compared markers between the corpora, while a qualitative thematic analysis explored expressions of self-stigma.

The quantitative analysis replicated the linguistic markers with culturally specific nuances. Compared to the control group, the depression corpus showed significantly elevated self-focus, negative emotions, and cognitive processing. A counterintuitive temporal pattern (past < present < future) was found, diverging from Western models of hopelessness. The qualitative analysis identified four themes of self-stigma: Internalized Shame, Cognitive Rigidity, Rumination and Hopelessness, and Cultural Intersections. These themes were linked to markers and often mitigated by self-rescue narratives in about 45 % of posts.

The findings extend Beck's cognitive theory and Corrigan's self-stigma model to digital, collectivistic contexts, highlighting social media's dual role in expressing distress and fostering resilience. The results have clear implications for designing culturally tailored mental health interventions and developing more sensitive AI-based tools for stigma and depression detection in non-Western populations.