Computational text analysis of emotional expressions related to non-suicidal self-injury in Chinese social media

Non-suicidal self-injury (NSSI) is closely associated with emotion dysregulation, yet emotional experiences may be underreported in traditional research due to cultural and social constraints. This study aimed to characterize emotional expressions related to NSSI in Chinese social media discourse. Public posts from Zhihu, Weibo, and Xiaohongshu (2019–2025) were screened, yielding 4,185 NSSI-related texts. A hybrid lexicon-based sentiment analysis classified emotions into seven categories, supplemented by unsupervised content clustering. Negative emotions comprised over 70% of expressions. Sadness was most prevalent, followed by fear, anger, and disgust. Sadness reflected chronic emotional invalidation; fear centered on disclosure and social consequences; anger was internalized and self-directed; disgust related to stigma and body alienation. Positive emotions were infrequent: happiness indicated transient relief after self-injury, acceptance emerged through peer validation, and surprise was linked to dissociative or impulsive episodes. NSSI-related emotions expressed online are predominantly negative, chronic, and socially embedded. Social media analysis provides a complementary approach for understanding emotional processes in NSSI and may inform culturally sensitive interventions.