13 November 2017

Language analysis to help predict psychosis onset

Language analysis to help predict psychosis onset

Artificial intelligence is being harnessed by researchers from Orygen, the National Centre of Excellence in Youth Mental Health, to identify speech patterns that might indicate increased risk of psychosis in help-seeking young people.

Around three in every 100 people will experience a psychotic disorder in their lives. Psychotic symptoms vary, but include detachment from reality, disorganised or confused thinking, hallucinations, and lack of motivation or emotionality. Finding effective methods of identifying young people at risk of psychosis can ensure treatments are provided early.

The Orygen researchers have partnered with US-based researchers and computer scientists from IBM to analyse language use and its ability to predict the development of psychotic symptoms.

Associate Professor Barnaby Nelson, who is leading the project, said similar methods had successfully been used on a much smaller scale to predict onset of psychotic disorders. The current project attempts to replicate this initial strong finding on a larger scale and relate the findings to a wider array of symptoms and disturbances.  

“Language offers privileged insight into the mind, and is the main means by which we understand others,” Associate Professor Nelson said.

“This approach holds great promise in helping us identify speech patterns that might signal emerging thought disorder, which is a prominent feature of psychosis,” he said.

The research, which is funded by the National Institutes of Health (NIH) in the US, will involve 450 young people from Melbourne, Toronto and New York. Three groups will be studied: healthy volunteers, people at high-risk of psychosis and people who have recently experienced psychosis.

People who are interested in participating in the study should contact Barnaby Nelson – barnaby.nelson@orygen.org.au