Speech Patterns May Predict Psychosis Risk
An automated speech analysis program accurately identified at-risk young people who developed psychosis over a two-and-a-half year period in a proof-of-concept study published in NPJ-Schizophrenia. Researchers from Columbia University Medical Center, New York State Psychiatric Institute, and the IBM T. J. Watson Research Center collaborated on a computer program that analyzed each patient’s emantic coherence and syntactic structure in open-ended, narrative interviews spanning two and a half years in which they described their subjective experiences.
Speech features that predicted psychosis onset included breaks in the flow of meaning from one sentence to the next and speech that was characterized by shorter phrases with less elaboration. The analysis correctly differentiated between the five individuals who later experienced a psychotic episode and the 29 who did not, suggesting that this could help to identify thought disorder years prior to the onset of psychosis. Next steps include additional studies in larger groups of at-risk individuals but the tool is extremely promising, inexpensive, portable, fast, and non-invasive.