Psychosis is a complicated disorder that has long been left to the expertise of psychiatrists. However, that appears to be changing. A computer program is now able to predict the disorder simply by observing language patterns, according to a new study.
Thirty-four young people underwent computer testing for their language patterns’ semantic coherence – the flow of meaning from one sentence to the next – in each person during an open-ended, narrative interview. The speech tests were held quarterly over a period of 2.5 years.
The computer program demonstrated a 100 percent success rate in differentiating the five people who eventually experienced a psychotic episode and the 29 who did not. Unlike the computer used in the experiment, humans only have a 79 percent success rate of predicting whether a person will develop psychosis, based on tracking their speech patterns in interviews.
The study describing the experiment, conducted by researchers at Columbia University, the New York State Psychiatric Institute, and the IBM T.J. Watson Research Center, was published in the journal Schizophrenia on Wednesday.
It explains the use of an automated speech analysis program which managed to correctly predict the young people who would develop psychosis.
The interviews with the participants of the experiment were transcribed and analyzed by the computer program for patterns of speech, including semantics (meaning) and syntax (structure).
This was done using an algorithm developed by the researchers, which rooted out “jarring disruptions” in otherwise ordinary speech.
The speech characteristics that predicted psychosis included breaks in the flow of meaning from one sentence to the next, and speech characterized by shorter phrases with less elaboration.
"When people speak, they can speak in short, simple sentences. Or they can speak in longer, more complex sentences, that have clauses added that further elaborate and describe the main idea," Guillermo Cecchi, a biometaphorical-computing researcher for IBM Research, told The Atlantic.
"The measures of complexity and coherence are separate and are not correlated with one another. However, simple syntax and semantic incoherence do tend to aggregate together in schizophrenia,” he continued.
During 45 minutes of interviewing, those who went on to develop psychosis had at least one occasion of a “jarring disruption” in meaning from one sentence to the next.
“As an interviewer, if my mind wandered briefly, I might miss it. But a computer would pick it up,” Cecci said.
The computer method could be able to identify thought disorder in its earliest form, years before the onset of psychosis. Thought disorder is a key component of schizophrenia, a type of psychosis.
The findings could also open up possibilities for intervention.
"If speech analyses could identify those people most likely to develop schizophrenia, this could allow for more targeted preventive treatment before the onset of psychosis, potentially delaying onset or reducing the severity of the symptoms which do develop," said Gillinder Bedi, an assistant professor of clinical psychology at Columbia University.
The researchers plan to attempt to replicate their findings using a larger cohort of at-risk youths.
About one percent of the population between 14 and 27 is considered to be at clinical high risk (CHR) for psychosis. Those individuals have symptoms such as unusual or tangential thinking, perceptual changes, and suspiciousness. About 20 percent will go on to experience a full-blown psychotic episode.