An auditory-based machine learning algorithm was able to identify children diagnosed with depression and anxiety with 80% accuracy after analyzing recordings of their speech. The algorithm identified eight audio features that signify a higher risk of depression. Of these, a lower pitch of voice, repeatable speech inflections and a higher pitch response to surprise stimuli, were more indicative of depression. Researchers hope to develop a smartphone app that records and analyzes speech immediately, helping to better detect children at risk of internalizing disorders.
Researchers report computer based analysis of speech transcripts from teens at risk of psychosis were able to predict, with 83% accuracy, who would develop symptoms within 2 years.
Computerized speech analysis provides a more accurate classification than clinical ratings for teens at risk of developing psychosis, a new study reports.