Artificial intelligence (AI) can be used to predict all kinds of things, such as shopping habits, the weather, topics that interest the user, and so on. Recently there was even AI that could predict when a terminal patient might die, which has the potential of making end-of-life care better. Now researchers have developed a new AI that can predict psychosis.

Developed by IBM’s Computational Psychiatry and Neuroimaging research team, the AI is used to predict the risk of a patient developing psychosis. This is done by analyzing the speech patterns of 59 individuals, which was then further broken down into different parts and scored based on how coherent the sentences were. This was based off research work that was done in 2015 in a separate study.

From the previous work, 19 participants developed psychosis in two years, while the remaining 40 did not. The model that the researchers used to analyze the speech patterns to make its prediction was found to have 83% accuracy. It also found that those who might be at risk of developing psychosis used fewer possessive pronouns when speaking, and constructed less coherent sentences.

According to study author Guillermo Cecchi, “We believe this is a significant step towards the goal of developing a tool for mental health practitioners, caregivers and patients themselves with a tool that can expand and multiply the reach of neuropsychiatric assessment outside of the clinic.”

Filed in Computers >Medical. Read more about , and .