Contents lists available at ScienceDirect Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad Research paper Linguistic analysis of autobiographical narratives in unipolar and bipolar mood disorders in light of multiple code theory R. Mariani a, ⁎ , M. Di Trani a , A. Negri b , R. Tambelli a a Department of Dynamic and Clinical Psychology, Sapienza University, Rome, Italy b Department of Human and Social Sciences, University of Bergamo, Bergamo, Italy ARTICLE INFO Keywords: Bipolar disorders Mood disorders Language of depression Computerized linguistic measures Referential Process Sensory-Somatic ABSTRACT Background: Discriminating bipolar disorder (BD) from unipolar disorder (UD) is crucial in diagnosing mood disorders. Neurophysiological studies have identified different correlates of emotional regulation in BD and UD. According to the Multiple Code Theory, bodily modifications relate to linguistic styles, as highlighted by studies on the language of depression. Our purpose is to verify the existence in the Italian language of linguistic features of depression differentiating BD from UD to provide tools for clinicians to use beyond self-report measures. Methods: The sample included 20 BD, 20 UD (all diagnosed using DSM-5), and 20 Control Group (CG) parti- cipants. Participants completed the Profile of Mood States (POMS) and an audio-recorded Relationship Anecdotes Paradigm Interview, transcribed and analyzed by the Discourse Attributes Analysis Program for Referential Process Linguistic Measures. Results: One-way ANOVAs confirmed that specific linguistic features characterized BD, UD and CG. The use of Sensory-Somatic words was significantly different in the groups: higher in BD, intermediate in UD, and lower in CG. Individuals with BD produced higher scores on the Referential Activity Intensity Index and the use of sin- gular pronoun “I”. Negative Affect, as well as several POMS subscales, distinguished UD and BD from CG. Limitations: Narrow sample size, use of a single self-report instrument and treatment effects on measures in the clinical groups are limitations of the study. Conclusion: Individuals with UD and BD appear to use sensory-somatic language in predictably different patterns from each other and from the non-clinical population. Observation and assessment of linguistic features could improve diagnostic accuracy. 1. Introduction 1.1. Relevance and difficulties in distinguishing unipolar and bipolar mood disorders Mood disorders are expected to figure as the second most prevalent diseases worldwide after cardiovascular pathologies in this decade. (GBD 2017 Disease and Injury Incidence and Prevalence Collaborators, 2018). Reliable assessment of the different mood dis- orders is, which is still challenging in many cases, is critical for plan- ning effective treatment. A major clinical factor that impacts patient care and safety is differentiating between bipolar disorder (BD) and unipolar disorder (UD). BD is a debilitating illness that seriously affects quality of life, impairs social functioning, and increases suicide risk. Even when an individual's mood disorder presents almost exclusively as a series of depressive episodes, the presence of any manic or hypomanic episodes, however brief, necessitates a diagnosis of BD (Dean et al., 2004; Grande et al., 2016; Wang et al., 2017). For this reason, a critical aspect of clinical assessment is correctly identifying BD during de- pressive periods. Further complicating the assessment is the recognition of the symptoms of mania, frequently occurring only in a mild form (Cardoso de Almeida and Phillips, 2013; Phillips and Kupfer, 2013). Self-report questionnaires are not always suitable for discriminating BD from UD, because patients demonstrate self-perception bias about their moods and states. In fact, both BD and UD patients tend to over/under- estimate feelings (Lam, Wright, & Sham, 2005; Alatiq et al., 2010). This self-report bias not only affects mood disorders but all psychopatholo- gies, since individuals are influenced by their altered perception and by the psychopathology itself (John and Robins, 1994; Goffin and Gellatly, 2001; Di Trani et al., 2018). As a result of these issues, BD misdiagnosis as UD is a common event affecting about 69% of patients, who generally meet with an average of four physicians before they are https://doi.org/10.1016/j.jad.2020.03.170 Received 1 October 2019; Received in revised form 21 February 2020; Accepted 29 March 2020 ⁎ Corresponding author. E-mail address: rachele.mariani@uniroma1.it (R. Mariani). Journal of Affective Disorders 273 (2020) 24–31 Available online 06 May 2020 0165-0327/ © 2020 Elsevier B.V. All rights reserved. T