Predicting treatment response in psychosis using fMRI: A comprehensive review

Area of Science:

• Neuroscience
• Psychiatry
• Medical Imaging

Background:

• Schizophrenia (SCZ) treatment response prediction remains a challenge.
• Functional Magnetic Resonance Imaging (fMRI) and machine learning (ML) are increasingly used to predict treatment outcomes.
• Personalized medicine approaches are needed to optimize psychosis treatment.

Purpose of the Study:

• To review the literature on fMRI measures for predicting pharmacological treatment response in psychosis.
• To assess the role of statistical and ML algorithms in treatment outcome prediction.
• To identify potential brain network biomarkers for personalized treatment.

Main Methods:

• Comprehensive literature review of fMRI studies predicting treatment response in psychosis.
• Searched PubMed from January 1990 to December 2023.
• Included 21 relevant fMRI studies.

Main Results:

• Machine learning (ML) techniques show potential for enhancing prediction accuracy.
• Resting-state fMRI identified associations between functional connectivity and treatment response.
• Altered connectivity in Default Mode Network (DMN), Salience Network (SN), Central Executive Network (CEN), and sensory-motor circuits was noted.

Conclusions:

• fMRI measures, particularly functional connectivity patterns, may offer predictive value for psychosis treatment response.
• Personalized treatment strategies could be informed by these neuroimaging findings.
• Further research is needed to develop robust and generalizable predictive models for psychosis.