(PREPRINT) Plank & Zlomuzica, 2024, coherence in social media posts

The extraction of linguistic markers from social media posts which are indicative of the onset and course of mental disorders offers great potential to mental healthcare. In the present study, we extracted over one million posts from the popular social media platform reddit to analyze speech coherence which reflects formal thought disorder and is a characteristic feature of schizophrenia and associated psychotic disorders. Natural language processing (NLP) models were used to perform an automated quantification of speech coherence. While controlling for the emotional tone and subjectivity, we demonstrated that posts made in forums geared towards disorders with a higher degree of psychotic and formal thought disorder symptoms tend to show a lower level of coherence. The lowest coherence scores were detected in posts made in forums on dissociative identity disorder, schizophrenia and bipolar disorder followed by a medium level of coherence in posts submitted to a forum on autism and post-traumatic stress disorder. In contrast, a relatively high level of coherence was detected in posts related to obsessive compulsive disorder, depression, anxiety and attention-deficit hyperactivity disorder. Our findings provide promising first evidence for a possible utility of NLP-based coherence analyses for the early detection and prevention of psychosis on the basis of posts gathered from publicly available social media data. This opens new avenues for large-scale prevention programs aimed at high-risk populations.