At Columbia University, a team of researchers used speech-analysis software to predict the onset of schizophrenia with 100% accuracy in a sample group of 34 young women.
It's common practice in psychiatry to look for predictive qualities in patterns of speech. Disjointed patterns of speech -- which are evidence of underlying confusion and loss of logical flow in a person's thoughts -- are often a sign that a patient will develop schizophrenia. In this way, doctors can predict whether a patient will experience the serious mental illness with 79% accuracy .
This study shows that computers are even better at analyzing patterns of speech and making these predictions. The sample size was small and further testing will have to be completed, but in this preliminary trial, out of the 34 at-risk women in the group, the speech-analysis software accurately identified all five who would go on to develop schizophrenia -- a huge leap for medical technology.