31 January 2018

Computerised speech analysis successfully predicts psychosis in at-risk patients

Computerised speech analysis successfully predicts psychosis in at-risk patients

A computer-based program that analyses speech has successfully predicted, with 83% accuracy, which young people at risk of psychosis would go on to develop the disorder within two years, a study has found.

The study, undertaken in collaboration with Orygen, The National Centre of Excellence in Youth Mental Health, and researchers in the US and Argentina, revealed the same program predicted the onset of psychosis in a second group of at-risk young people with an accuracy of 79%.

Orygen’s Dr Gill Bedi said the study’s findings suggest that language-classifying technology has the potential to improve prediction of psychosis among young people at ultra-high risk of developing the disorder. Around three in every 100 young people will experience a psychotic episode, with most fully recovering from the experience.

“One of the characteristics of psychosis is disorganised thinking, which can present as a disrupted flow in people’s speech where their conversation seems to go off track,” Dr Bedi said.  

 “Having this technology could really help young people get the treatment they need earlier and ultimately alter the way we do psychiatry and potentially other fields of medicine,” she said.

The results of the study are published in the current issue of the journal World Psychiatry.

Dr Bedi said the next stage of the research would focus on testing the effectiveness of the language-classifying technology in larger groups of people – 93 young people were involved in the current study – and adapting it to analyse languages other than English.

This research was funded by the US National Institute of Mental Health.