Hotbed of stigmatization or source of support: A multimodal analysis of mental health-related videos on Douyin

Abstract

This study investigates how mental health stigma is constructed and contested through multimodal discourses on Douyin, one of China's most influential short-video platforms. Drawing on a stratified sample of 200 high-engagement videos posted between 2018 and 2023 across 20 widely viewed mental health–related hashtags, we extract and analyze 3416 image frames using a multimodal social semiotic framework. Integrating computer vision techniques, AI-assisted annotation, and moral foundations analysis, we examine how visual, textual, and auditory elements coalesce to encode stigma or promote anti-stigma narratives. Stigmatizing content frequently employs dark color schemes, high-angle shots, and emotionally charged audio to evoke fear and reinforce moral framings of degradation. In contrast, anti-stigma videos tend to feature warmer aesthetics and narratives grounded in appeals to care, fairness, and resilience. Discursive patterns also vary by diagnosis, reflecting distinct rhetorical and affective framings. Theoretically, this study reconceptualizes mental health stigma as a visually mediated and morally framed phenomenon shaped by the logics of algorithmic platforms. Douyin thus functions as both a vector of stigmatization and a potential site for empathetic, inclusive discourse, offering practical implications for digital mental health advocacy in China.