EDMONTON -- Researchers at the University of Alberta are developing a tool that they say can predict schizophrenia using artificial intelligence.

The machine learning tool analyzes brain scans to identify the risk of a schizophrenia diagnosis.

In a recent study, it was used to analyze MRI images of 57 people who were either the siblings or children of schizophrenia patients. The results, published in Nature Partner Journals, show that it accurately identified the people who scored the highest on a self-reported personality trait scale.

“The goal is for the tool to help with early diagnosis, to study the disease process of schizophrenia and to help identify symptom clusters,” said lead author Sunil Kalmady Vasu.

Schizophrenia can cause hallucinations, disorganized speech and trouble thinking. First-degree relatives of people diagnosed with schizophrenia have a higher risk of developing it in their lifetime.

Kalmady Vasu, a senior machine learning specialist with the U of A's Faculty of Medicine & Dentistry and a member of the Alberta Machine Intelligence Institute, says the tool is not intended to replace a diagnosis by a psychiatrist. But it could be more accurate than self-assessment.

The tool, called EMPaSchiz (Ensemble algorithm with Multiple Parcellations for Schizophrenia prediction), will next be tested on people with no family connection to schizophrenia.

It will also be used to track assessed individuals to see if they will develop schizophrenia.

EMPaSchiz was developed by researchers from the U of A and the National Institute of Mental Health and Neurosciences in India.