"Unmasking Misinformation": Characterizing the Social Media Landscape of Disorders of Consciousness on X (formerly...

Objective To characterize the discursive landscape of public discussion related to disorders of consciousness (DoC) on the X social media platform, including its nature, prevalence, thematic structure, and contextual diversity. Methods Point-prevalence, observational study of social media content using a combined social network analysis and unsupervised topic-modeling approach. The setting was X (formerly Twitter), a global public microblogging platform. English-language posts published between January 1 and December 31, 2023 were retrieved using a curated set of nine DoC-related keyword phrases ("Disorders of Consciousness," "Coma Recovery," "Vegetative State," "Minimally Conscious State," "Brain Death," "Locked-In Syndrome," "Medically Induced Coma," "Coma Prognosis," "Dead Zone"). No interventions were applied. Results A total of 30,621 posts were retrieved. Social network analysis using the Harel-Koren Fast Multiscale layout and Clauset-Newman-Moore modularity-based clustering identified seven principal user communities. BERTopic transformer-based topic modeling yielded 50 coherent topics, which were grouped through structured thematic synthesis into three dominant discourse domains: (1) medical and clinical (vegetative state, brain death, locked-in syndrome, medically induced coma, and prominent named cases); (2) ethical, legal, and societal (organ donation/procurement, end-of-life and beginning-of-life debates, criminal justice cases involving vegetative outcomes); and (3) figurative and non-medical (the term "dead zone" used in technology, gaming, popular culture, and Stephen King's novel). Non-medical uses constituted a substantial fraction of total retrieval volume, and the unsupervised model was unable to algorithmically separate clinical from figurative uses of overlapping terms. Conclusions Public discourse about DoC on X is substantial in volume, thematically heterogeneous, and embedded within emotionally charged, legally complex, and figurative contexts that dilute clinical meaning and create fertile conditions for misinformation. These findings provide an evidence-based map of where targeted educational and public-health interventions are most needed and underscore the value of cross-platform, multilingual, and domain-specific NLP follow-up work.