Subcortical and insula functional connectivity aberrations and clinical implications in first-episode schizophrenia

Abstract

Introduction

Schizophrenia is a complex mental disorder whose pathophysiology remains elusive, particularly in the roles of subcortex. This study aims to explore the role of subcortex and insula and their relationship with symptom changes in first-episode schizophrenia (FES) patients by utilizing machine learning algorithms and functional connectivity (FC).

Methods

The study encompasses 261 participants, sourced from two independent samples of FES patients and their matched healthy controls (HC). The discovery dataset includes 77 FES patients at baseline (FES0W) and 77 matched HCs, with the patients undergoing a follow-up scan after eight weeks of antipsychotic treatment (FES8W, N = 34). A validation dataset from another region comprises 47 FES patients and 47 matched HCs.

Results

Significant differences in subcortical FCs were observed between FES and controls, correlating with symptom severity and symptom changes. Machine learning models were developed to diagnose schizophrenia on an individual basis, achieving a balanced accuracy of 79.55 % across diverse centers.

Conclusions

These findings suggest that subcortical connectivity patterns offer potential as biomarkers for schizophrenia, enabling personalized treatment strategies and improving prognosis by facilitating early diagnosis.