Researchers at the Royal College of Surgeons in Ireland (RCSI) have combined a blood test with machine learning to predict which individuals are likely to face severe psychotic disorders in their futures. The work could help head off such problems.
The work focused on people who are known to be at high risk of psychosis—generally those who have exhibited mild or temporary psychotic behavior in the past. The researchers analyzed blood samples taken from 133 such individuals participating in two ongoing health-study programs, and then examined them several years later to see if they had developed a psychotic disorder, such as schizophrenia.
Predicting the Future
With this knowledge, the researchers then looked for current markers in the blood that might be used to predict future psychotic disorders. Using machine learning to analyze the data, they found ten key proteins that seemed to indicate whether or not someone at high risk of developing a psychotic disorder would, in fact, do so. The program was then able to predict the incidence of future psychosis with a 93% accuracy. It was also able to predict the absence of future disorders 80% of the time.
Interestingly, many of the protein markers linked to psychosis have also previously been linked to inflammation, leading the researchers to believe that early changes in the immune system—which triggers inflammation to fight disease—might be linked to the propensity to develop a psychotic disorder.
“Ideally, we would like to prevent psychotic disorders, but that requires being able to accurately identify who is most at risk,” said Professor David Cotter, the study’s senior and corresponding author and professor of molecular psychiatry at RCSI. “Our research has shown that, with help from machine learning, analysis of protein levels in blood samples can predict who is at truly at risk and could benefit from preventive treatments. We now need to study these markers in other people at high risk of psychosis to confirm these findings.”
The initial study has been published in the peer-reviewed journal, JAMA Psychiatry.