Brainwide Measures of Functional Coupling May Not Be Suitable for Predicting Clinical Outcomes in Patients with Psychosis

Strategic Insights -

Research conducted by a team of scientists at Orygen, led by Sidhant Chopra, explored the potential of brainwide measures of functional coupling in predicting clinical outcomes for patients experiencing their first episode of psychosis (FEP). The study, published in Biological Psychiatry Global Open Science, aimed to identify prognostic biomarkers that could facilitate personalized treatments. Fifty-five antipsychotic-naive patients with FEP were recruited and underwent functional magnetic resonance imaging (fMRI) at baseline and after 3 months. The researchers analyzed the data using three different cross-validated prediction algorithms and found that brainwide measures of functional coupling were not suitable for predicting extended clinical outcomes over a 6- to 12-month period.

Key Takeaways:

  • The study aimed to identify prognostic biomarkers for patients experiencing their first episode of psychosis (FEP) to facilitate personalized treatments.
  • Fifty-five antipsychotic-naive patients with FEP were recruited and underwent functional magnetic resonance imaging (fMRI) at baseline and after 3 months.
  • The researchers used three different cross-validated prediction algorithms, including connectome-based predictive modeling, kernel ridge regression, and multilayer meta-matching.
  • Each prediction model comprised 35 to 49 individuals, but all models showed poor performance in predicting patients' 6- and 12-month changes in symptoms and functioning (all p > .05).
  • The study suggests that brainwide measures of functional coupling may not be suitable for predicting extended clinical outcomes over a 6- to 12-month period in patients with FEP.

Statistics:

  • 55 patients with FEP were recruited for the study (51% female, ages 15-25 years).
  • fMRI was conducted at baseline and after 3 months to evaluate the effect of antipsychotic medication on brain function.
  • The researchers considered 3 different prediction algorithms: connectome-based predictive modeling, kernel ridge regression, and multilayer meta-matching.
  • Each prediction model comprised 35 to 49 individuals.
  • The p-value for all prediction models was > .05, indicating poor performance in predicting patients' 6- and 12-month changes in symptoms and functioning.

Sources:

  • Chopra, S., Pope, I. Z., Holmes, A., Francey, S. M., O'Donoghue, B., Cropley, V. L., ... & McGorry, P. D. (2025). Functional Coupling and Longitudinal Outcome Prediction in First-Episode Psychosis. Biological Psychiatry Global Open Science, 5(6), 100589.
  • NewsRx. Data on Personalized Medicine Reported by Sidhant Chopra and Co-Researchers (Functional Coupling and Longitudinal Outcome Prediction in First-Episode Psychosis). Mental Health Weekly Digest. October 20, 2025; p 176.