Cloud Hosting

Machine Learning Tool Predicts Onset of Psychosis

A ground-breaking development in mental health research has emerged with the creation of a machine learning tool that can predict the onset of psychosis. The tool, developed by researchers from the University of Tokyo and an international consortium, classifies MRI brain scans into two groups: healthy individuals and those at risk of a psychotic episode.

The study, which involved over 2,000 participants from 21 global locations, has shown promising results. The tool was found to be 85% accurate at distinguishing between people who were not at risk and those who later displayed psychotic symptoms. Published in the journal Molecular Psychiatry, this breakthrough could lead to early intervention and improved clinical outcomes.

Machine learning approaches using structural magnetic resonance imaging (sMRI) are instrumental in disease classification and predicting psychosis onset in individuals at clinical high risk (CHR). The researchers created a model to differentiate between CHR individuals who later developed psychosis, healthy controls, and those with uncertain follow-up status.

Regional cortical surface area measures strongly contributed to the classification of CHR individuals. Structural MRI studies have shown alterations in brain structure among CHR individuals, including a progressive decrease in grey matter volume and changes in cortical surface area and thickness. These alterations suggest that baseline MRI scans for CHR individuals may be helpful in identifying their prognosis.

Early intervention based on the machine learning tools predictions can significantly improve outcomes for individuals at risk. The classifier was 85% accurate on the training set and 73% accurate on independent confirmatory datasets. The CHR state is widely used for early detection and prevention of psychotic disorders, and individuals at CHR have a higher risk of developing psychosis compared to healthy controls.

Machine learning combined with structural MRI scans can help identify biomarkers and patterns associated with psychosis onset. The findings are promising, suggesting that machine learning approaches could play a crucial role in predicting the onset of psychosis in high-risk individuals, leading to early intervention and more effective treatment strategies.

The research team is working on creating a more robust classifier for new data sets, and further prospective studies are required to determine whether the classifier could be helpful in clinical settings. Ultimately, the goal is to enable earlier intervention and targeted care for at-risk individuals, potentially revolutionizing mental health research and treatment.

Original post:
Machine Learning Tool Predicts Onset of Psychosis | Health News - Medriva