Researchers at the University of Alberta have taken a step further in developing artificial intelligence tools for predicting schizophrenia by analyzing brain scans.
A recently published study used this tool to analyze functional magnetic resonance imaging of 57 healthy first-degree relatives (brothers or children) in schizophrenic patients. We have accurately identified the 14 individuals who scored the highest on the self-reported schizophrenic personality trait scale.
Schizophrenia, which affects 300,000 Canadians, can cause delusions, hallucinations, chaotic speech, thinking problems, and lack of motivation, usually treated with a combination of drugs, psychotherapy, and brain stimulation. Will be done. Patients’ first-degree relatives have a general population risk of less than 1%, while a lifetime risk of developing schizophrenia is up to 19%.
“Our evidence-based tools can be more accurate than diagnosing by examining neural signs in the brain and subjectively assessing symptoms,” said Sunil Kalmady Vasu, senior author of machine learning specialists at the School of Medicine. Mr. says. Dentistry.
Kalmady Vasu said the tool was designed as a decision support tool and does not replace the diagnosis by a psychiatrist. He also pointed out that having schizophrenia personality traits may make people more vulnerable to mental illness, but it is uncertain whether they will develop full-blown schizophrenia.
“The goal is a tool that helps early diagnosis, study the disease process of schizophrenia, and identify symptom clusters,” said Karmady Vasu, a member of the Alberta Machine Intelligence Institute.
This tool, called EMPaSchiz (an ensemble algorithm with multiple perserations for predicting schizophrenia), previously predicted a diagnosis of schizophrenia with 87% accuracy by examining a patient’s brain scan. Was used for It was developed by a team of researchers at the University of Alberta and the National Institute of Psychiatry and Neuroscience in India. The team also includes three members of the University of Arizona’s Institute of Neuroscience and Mental Health. Computational scientist and CIFARAI Chairman of the Faculty of Science, Russ Greiner, and psychiatrists Andrew Greenshaw and Serdar Dursun, who are also authors of the latest treatise.
According to Kalmady Vasu, the next step in the study is to test the accuracy of the tool in nonfamilial individuals with schizophrenia characteristics, track the evaluated individuals, and develop schizophrenia later in life. Learn whether to do it.
Kalmady Vasu is also using the same principles to develop algorithms through the Canadian VIGOR Center to predict outcomes such as cardiovascular mortality and readmission for heart failure.
“Severe mental illness and cardiovascular problems cause dysfunction and impair quality of life,” said Karmadi Bass. “It is very important to develop objective, evidence-based tools for these complex obstacles that plague humanity.”
Improved AI-based tools improve the accuracy of schizophrenia diagnosis
Sunil Vasu Kalmady et al, Extending the schizophrenia diagnostic model to predict schizotypy of first-degree relatives, npj schizophrenia (2020). DOI: 10.1038 / s41537-020-00119-y
Courtesy of the University of Alberta School of Medicine and Dentistry
Quote: The AI used to predict the initial symptoms of schizophrenia in the patient’s relatives (January 26, 2021) is https: //medicalxpress.com/news/2021-01-ai-early-symptoms- Obtained from schizophrenia-relatives.html on January 26, 2021
This document is subject to copyright. No part may be reproduced without written permission, except for fair transactions for personal investigation or research purposes. The content is provided for informational purposes only.
AI used to predict early symptoms of schizophrenia in a patient’s relatives
Source link AI used to predict early symptoms of schizophrenia in a patient’s relatives