http://www.digitaljournal.com/tech-and-science/science/speech-analysis-software-predicts-psychosis/article/512923

Speech analysis software predicts psychosis

Posted Jan 23, 2018 by Tim Sandle
A new study shows how speech analysis software can predict a mental illness like psychosis, in at-risk patients, with up to 83 percent accuracy. This offers a potential diagnostic tool for psychiatry.
The research came from the The Mount Sinai Hospital and Mount Sinai School of Medicine, and it signals how big-data approaches have potential to improve prediction of psychiatric along with other medical disorders. This was assessed via computer-based analyses of speech transcripts which were obtained from interviews with at-risk youths.
The computer analysis was able to predict which youths would later go onto develop psychosis within a two year framework (with an accuracy of 83 percent). This was demonstrated using two independent cohorts of young people determined by medical doctors to be at risk for psychosis.
With psychosis there is a disturbance in the flow of meaning when speaking (this is sometimes referred to as being tangential or going off track). The analysis of the speech patterns was performed in order to predict which subjects would later develop psychosis. Psychosis is medically defined as an abnormal condition of the mind that involves a loss of contact with reality. People experiencing psychosis often exhibit personality changes and thought disorder. Here disorganized thinking is often evidenced by disorganized speech.
The research considered transcripts from interviews undertaken with at-risk young people from two different cohorts. The first one was from New York City with 34 participants and the second was from Los Angeles with 59 participants. The transcripts were analyzed by computer deploying automated natural language processing methods. This approach was used to assess differences in speech between those who developed psychosis and those who did not.
According to lead researcher Dr. Cheryl Corcoran, who is quoted by PsychCentral: "The results of this study are exciting because this technology has the potential to improve pre-diction of psychosis and ultimately help us prevent psychosis by helping researchers develop re-mediation and training strategies that target the cognitive deficits that may underlie language disturbance."
The findings could lead to the development of novel computerized methods to characterize and predict earlier complex behaviors, like psychosis. The research is published in the journal World Psychiatry. The peer reviewed study is headed "Prediction of psychosis across protocols and risk cohorts using automated language analysis."