Content-free speech activity records: interviews with people with schizophrenia

Abstract Schizophrenia is characterised by a variety of symptoms, many which are expressed verbally. However, privacy concerns limit the possibility collecting and sharing large corpora schizophrenic speech. As result, variability in communicative behaviour individuals with schizophrenia currently poorly understood. In this study we explore how far content-free speech activity records can successfully profile patients schizophrenia. We used data from one very few publicly available conversations schizophrenia, featuring interviews between therapist three different symptoms. Crucially, annotated only moments silence during interview, entirely discarded verbal content interview. way perform type analysis that fully preserves speakers’ privacy, while still allowing for comparisons full original recorded material. developed several types analyses corresponding visualisations records. Exemplifying these analyses, clear patient-specific profiles be derived, based on dimensions total duration therapists. These findings consistent an independent phenomenological assessment patients.