Linguistic Markers of Anorexia Nervosa: Preliminary Data from a Prospective Observational Study

Recent works indicated the potential relevance of Natural Language Processing techniques for the detection of clinical conditions. This paper tries to address the issue in the Eating Disorder domain, by exploiting “linguistic biomarkers” for Anorexia Nervosa (AN) detection in female teenagers. We hypothesize that (i) disturbances in self-perceived body image, black and white thinking and mood changes strongly associated with AN disorder can result in altered linguistic patterns; and (ii) these subtle modifications can be identified by means of NLP tools, acting as early proxy measures for the disorder. To this aim, we enrolled 51 participants (age range: 14-18): 17 girls with a clinical diagnosis of Anorexia Nervosa and 34 normal weighted peers, matched by gender, age and educational level. Both the groups were asked to produce three written texts (around 10-15 lines long), i.e. two autobiographical narratives and a short description of a complex figure. A rich set of linguistic features was extracted from the text samples and the statistical significance in pinpointing the pathological process was measured. Our preliminary results show that subtle language disruptions, mainly at the lexical and syntactic level, can actually represent an early but reliable marker of the disease. However, an analysis on a bigger cohort with follow-up information, still ongoing, is needed to consolidate this assumption.