Predictive Linguistic Features of Schizophrenia

Proceedings of the 6th Joint Conference on Lexical and Computational Semantics (*SEM 2017), pages 241–250, Vancouver, Canada, August 3-4, 2017. c⃝2017 Association for Computational Linguistics Predictive Linguistic Features of Schizophrenia Efsun Sarioglu Kayi1, Mona Diab1, Luca Pauselli2, Michael Compton2, Glen Coppersmith3 1Department of Computer Science, George Washington University {efsun, mtdiab}@gwu.edu 2Medical Center, Columbia University {mtc2176@cumc.columbia.edu, pausell@nyspi.columbia.edu} 3Qntfy glen@qntfy.com Abstract Schizophrenia is one of the most disabling and difficult to treat of all human medi- cal/health conditions, ranking in the top ten causes of disability worldwide. has been a puzzle in part due to diffi-It culty in identifying its basic, fundamental components. Several studies have shown that some manifestations of schizophrenia (e.g., the negative symptoms that include blunting of speech prosody, as well as the disorganization symptoms that lead to dis- ordered language) can be understood from the perspective of linguistics. However, schizophrenia research has not kept pace with technologies in computational lin- guistics, especially in semantics and prag- matics. As such, we examine the writ- ings of schizophrenia patients analyzing their syntax, semantics and pragmatics. In addition, we analyze tweets of (self pro- claimed) schizophrenia patients who pub- licly discuss their diagnoses. For writ- ing samples dataset, syntactic features are found to be the most successful in classifi- cation whereas for the less structured Twit- ter dataset, a combination of features per- formed the best. 1 Introduction Schizophrenia is an etiologically complex, hetero- geneous, and chronic disorder. It imposes major impairments on affected individuals, can be devas- tating to families, and it diminishes the productiv- ity of communities. Furthermore, schizophrenia is associated with remarkably high direct and indi- rect health care costs. Persons with schizophrenia often have multiple medical comorbidities, have a tragically reduced life expectancy, and are of- ten treated without the benefits of sophisticated measurement-based care. Similar to other psychoses, schizophrenia has been studied extensively on the neurological and behavioral levels. Covington et al. (2005) note the existence of many language abnormalities (in syntactic, semantic, pragmatic, and phonetic do- mains of linguistics) comparing patients to con- trols. They observed the following: • reduction in syntax complexity (Fraser et al., 1986); • impaired semantics, such as the organization of individual propositions into larger struc- tures (Rodriguez-Ferrera et al., 2001); • abnormalities in pragmatics which is a level obviously disordered in schizophrenia (Cov- ington et al., 2005); • phonetic anomalies like flattened intonation (aprosody), more pauses, and constricted pitch/timbre (Stein, 1993). A few studies have used computational meth- ods to assess acoustic parameters (e.g., pauses, prosody) that correlate with negative symptoms, but schizophrenia research has not kept pace with technologies in computational linguistics, espe- cially in semantics and pragmatics. Accordingly, we analyze the predictive power of linguistic fea- tures in a comprehensive manner by computing and analyzing many syntactic, semantic and prag- matic features. This sort of analysis is particularly useful for finding meaningful signals that help us better understand the mental health conditions. To this end, we compute part-of-speech (POS) tags and dependency parses to capture the syntactic in- …

Content truncated to 500 words