Hallucinated and spoken linguistic patterns as markers of psychiatric disorders

A person with alterations in the brain and cognitive functioning and whose language- and speech-related processes are affected might experience atypical linguistic phenomena. Hallucinated voices, also known as auditory verbal hallucinations, illustrate this: when they manifest, an individual can hear words, phrases or dialogues that can resemble actual human language production in the absence of an actual source of the voice in the outer world. Disorganized speech represents another example: an individual might express her/himself with a discourse whose associations of concepts and use of grammatical elements area typical, sometimes until the point in which the speech is not understandable anymore. Since these linguistic phenomena occur both in individuals with and without need for mental care, a main clinical problem consist in accurately and consistently identifying the patterns that differentiate between pathological and non-pathological hallucinated voices or disorganized speech. In this thesis, a combination of linguistic theory, computational methods, and artificial intelligence was implemented to unveil the linguistic patterns that distinguish between pathological and non-pathological hallucinated voices, as well as those that differentiate the speech of patients with schizophrenia-spectrum disorders from that of control individuals. More broadly, a consensual framework was developed regarding the potential use of this approach across psychiatric disorders and for a series of clinical actions. Lastly, pending obstacles and emerging questions related to this approach were underlined, followed by tentative theoretical venues that might point into solutions and answers about “Hallucinated and spoken linguistic patterns as markers of psychiatric disorders”.