Cognitive-Linguistic Network Structure of OCD Personal Narratives: A Three-Group Comparison Using TEA Networks and Recurrence Quantification Analysis

Personal narratives encode cognitive patterns that vary across mental health conditions, but few studies have compared how different disorders shape the structure of self-narration. This study uses two computational methods, Target-Event-Agent (TEA) cognitive networks and Recurrence Quantification Analysis (RQA), to compare personal narratives of obsessive-compulsive disorder (OCD) with those of physical illness (cancer) and other mental health conditions (depression/PTSD). We analyzed 152 OCD recovery stories from a blog, 152 cancer patient stories from an interview-based website, and 200 depression/PTSD posts from Reddit. OCD narratives sit between physical illness and affective disorder narratives on both self-focused agency and thematic recurrence. After Benjamini-Hochberg correction, 16 of 19 originally significant effects survived at FDR = .05. Supplementary analyses showed that the gradient was driven primarily by depression (g = -0.71, p < .001), with a smaller, non-significant gap for PTSD (g = -0.32, p = .11)