A predictive model for negative and positive symptoms of schizophrenia based on resting-state fMRI

In recent years, the relationship between the positive and negative symptoms of schizophrenia and the brain's functional network has garnered increasing attention. In this study, 50 schizophrenia patients and 50 age- and gender-matched healthy controls were recruited. Resting-state fMRI data were acquired, and functional connectivity networks were constructed using Pearson correlation. To comprehensively investigate alterations in brain functional networks, six topological metrics were analyzed: degree, clustering coefficient, small-worldness, shortest path length, local efficiency, and global efficiency. Multiple linear regression models were developed to predict clinical symptom severity based on these network features. Compared to healthy controls, schizophrenia patients exhibited significantly reduced clustering coefficient and shortest path length, while small-worldness increased. Furthermore, machine learning-based prediction revealed that the clustering coefficient was the most effective predictor of negative symptoms (SANS score), whereas local efficiency best predicted positive symptoms (SAPS score). These findings provide new insights into the neural underpinnings of schizophrenia and offer potential biomarkers for its diagnosis and prognosis.