Scientists utilizing artificial intelligence have discovered a hidden clue in individuals’ language that may precisely predict whether or not they’re likely to develop psychosis in the future.
The machine-learning method more exactly quantifies the semantic richness of individuals’ conversational language, a recognized indicator for psychosis.
The analysis, revealed within the journal npj Schizophrenia, reveals that automated analysis of the two language variables―more frequent use of phrases associated with sound and talking with low semantic density, or vagueness―can predict whether an at-risk individual will later develop psychosis with 93 % accuracy.
Even educated clinicians had not noticed how folks in danger for psychosis use more words associated with sound than the average, though the abnormal auditory perception is a pre-medical symptom.
“Attempting to listen to these subtleties in conversations with individuals is like trying to see microscopic germs with your eyes,” stated Neguine Rezaii, who carried out the analysis at Emory University within the US.
“The automated method we have developed is a delicate tool to detect these hidden patterns. It is like a microscope for warning signs of psychosis,” stated Rezaii, who’s now at Harvard University within the US.
“It was previously recognized that subtle features of future psychosis are present in people’s language. However, we have used machine learning to uncover hidden details about those features,” mentioned Phillip Wolff, a professor at Emory University.
The findings add to the proof displaying the potential for utilizing machine learning to identify linguistic abnormalities related to mental illness, said Elaine Walker, an Emory professor.