Quantified speech as a marker: computational linguistics in schizophrenia-spectrum disorders

Speech contains a wealth of information about the speaker's state of mind, not only in the words used, but also in the way these words are pronounced. Recent developments in Natural Language Processing (NLP) have paved the way for rapid, systematic recording and analysis of quantifiable properties of spoken language. Schizophrenia spectrum disorders, a collection of serious psychiatric disorders, involve different aspects of language. Abnormalities in language are used by practitioners to make a diagnosis or to estimate the severity of complaints. By using NLP techniques it has become possible to make these deviations objective and quantifiable, so that subtle changes can also be detected. This thesis contains a number of studies on the application of NLP techniques to the spoken language of people with a schizophrenia spectrum disorder. The focus of the research presented here is mainly on semantics (what is told) and phonetics of speech (how is told). These aspects are used to distinguish individuals with schizophrenia spectrum disorders from healthy controls, to investigate the relationship between speech features and specific symptoms, the relationship with clinical subtypes, as well as the relationship with antipsychotic medication.