New computer based study identifies patients who are at risk of developing psychosis
Researchers from Columbia University Medical Center, New York State Psychiatric Institute and the IBM TJ Watson Research Center found that data obtained from computer speech-analysing can be more accurate than clinical studies, in identifying people with high risk of developing psychosis.
Computer speech-analysing objective data could provide important additional clinical tools that may be very helpful in diagnosing, prognosis and monitoring severe psychiatric disorders.
“Computerized analysis of complex human behaviours such as speech may present an opportunity to move psychiatry beyond reliance on self-report and clinical observation toward more objective measures of health and illness in the individual patient”, researchers say.
Speech offers “important clues about what people are thinking and feeling.” Psychiatry could find great use in speech-analysing computers. The method is more accurate in measuring different speech variables. This offers more reliable results when compared to the clinicians assessments based on intuition.
The computer analysed semantics and syntax of participants subjective narrated experiences. Coherence and phrase length were also used as indicators.
Study participants were then assessed every 3 months for the following 2.5 years in the attempt to predict their behaviour. “Classification based on automated analysis outperformed that based on clinical ratings, indicating that automated speech analysis can increase predictive power beyond expert clinical opinion”, authors observed.
Shorter, less elaborated phrases, lack of consistency and coherence were found to be linked with further development of psychotic episodes. Because computer speech-analysing data was found to be 100% accurate, the next step for the researchers is to use the method for identifying even earlier schizophrenia signs, the ones that develop before psychosis does.