In the Journals

Computer-based analysis of language may predict psychosis onset

A computer-based natural language processing analyses of speech transcripts predicted — with an 83% accuracy — young people at risk for developing later psychosis, researchers found.

“Language disturbance is prevalent in schizophrenia and is related to functional disability, given that an individual needs to think and speak clearly in order to maintain friends and a job,” Cheryl M. Corcoran, MD, from the department of psychiatry, Icahn School of Medicine at Mount Sinai, and colleagues wrote. “Beyond clinical ratings, there has been an effort to characterize early subtle language disturbances in clinical high-risk individuals using linguistic analysis, with the aim of improving prediction. While this manual linguistic approach appears to be superior to clinical ratings in psychosis prediction, it depends on predefined measures that may not capture other subtle language features.”

In a previous study, Corcoran and colleagues used computer-based natural language processing analyses among English-speaking young people at high risk for mental disorders and found that reduction in semantic coherence and in syntactic complexity predicted later psychosis onset with high accuracy. In this study, the researchers used these same methods in another, larger-risk cohort to cross-validate the effectiveness of an automated machine-learning speech classifier in predicting psychosis. They examined transcripts from 1-hour interviews with 94 high-risk youths to determine differences in speech between those who developed psychosis within 2 years and those who did not.

The automated machine-learning speech classifier had an 83% accuracy in predicting psychosis onset in high-risk young people. When using this method to cross-validate their previous findings, the software showed 79% accuracy in predicting psychosis onset in the original-risk cohort. In addition, it showed 72% accuracy in separating the speech of recent-onset psychosis patients from that of their healthy counterparts. The natural language processing approach was highly correlated with previously identified predictive automated and manual linguistic features, indicating its concurrent validity.

"The results of this study are exciting because this technology has the potential to improve prediction of psychosis and ultimately help us prevent psychosis by helping researchers develop remediation and training strategies that target the cognitive deficits that may underlie language disturbance," Corcoran said in a press release. "More broadly, language and behavior are the primary sources of data for psychiatrists to diagnose and treat mental disorders. There are now novel computerized methods to characterize complex behaviors such as language. Speech is easy to collect and inexpensive to analyze using computer-based analysis. This technology could be applied across psychiatry, and plausibly in other fields of medicine." – by Savannah Demko

Disclosures: The authors report no relevant financial disclosures.

A computer-based natural language processing analyses of speech transcripts predicted — with an 83% accuracy — young people at risk for developing later psychosis, researchers found.

“Language disturbance is prevalent in schizophrenia and is related to functional disability, given that an individual needs to think and speak clearly in order to maintain friends and a job,” Cheryl M. Corcoran, MD, from the department of psychiatry, Icahn School of Medicine at Mount Sinai, and colleagues wrote. “Beyond clinical ratings, there has been an effort to characterize early subtle language disturbances in clinical high-risk individuals using linguistic analysis, with the aim of improving prediction. While this manual linguistic approach appears to be superior to clinical ratings in psychosis prediction, it depends on predefined measures that may not capture other subtle language features.”

In a previous study, Corcoran and colleagues used computer-based natural language processing analyses among English-speaking young people at high risk for mental disorders and found that reduction in semantic coherence and in syntactic complexity predicted later psychosis onset with high accuracy. In this study, the researchers used these same methods in another, larger-risk cohort to cross-validate the effectiveness of an automated machine-learning speech classifier in predicting psychosis. They examined transcripts from 1-hour interviews with 94 high-risk youths to determine differences in speech between those who developed psychosis within 2 years and those who did not.

The automated machine-learning speech classifier had an 83% accuracy in predicting psychosis onset in high-risk young people. When using this method to cross-validate their previous findings, the software showed 79% accuracy in predicting psychosis onset in the original-risk cohort. In addition, it showed 72% accuracy in separating the speech of recent-onset psychosis patients from that of their healthy counterparts. The natural language processing approach was highly correlated with previously identified predictive automated and manual linguistic features, indicating its concurrent validity.

"The results of this study are exciting because this technology has the potential to improve prediction of psychosis and ultimately help us prevent psychosis by helping researchers develop remediation and training strategies that target the cognitive deficits that may underlie language disturbance," Corcoran said in a press release. "More broadly, language and behavior are the primary sources of data for psychiatrists to diagnose and treat mental disorders. There are now novel computerized methods to characterize complex behaviors such as language. Speech is easy to collect and inexpensive to analyze using computer-based analysis. This technology could be applied across psychiatry, and plausibly in other fields of medicine." – by Savannah Demko

Disclosures: The authors report no relevant financial disclosures.