What Static Connectivity Misses : Dynamic Alpha-Band Brain Network States in First-Episode Psychosis

Abstract

This study investigates resting-state EEG alpha-band connectivity in first-episode psychosis (FEP) using a novel dynamic connectivity pipeline and examines its relationship with cognitive functioning and psychopathological scores. Data from 78 individuals with FEP and 60 healthy controls (CTR) were analyzed. Source estimation was performed using eLORETA, and connectivity was quantified with the weighted phase-lag index. Static connectivity matrices were assessed using graph theory and edge-wise metrics. Dynamic connectivity matrices were clustered into five distinct brain network states (BNS) using a modified k-means algorithm, from which temporal and graph theory metrics were extracted. Static connectivity metrics revealed lower alpha connectivity in FEP than in controls. The dynamic approach identified reduced variability in the characteristic path length within the default mode network–associated BNS 1 in FEP, suggesting diminished adaptive modulation of functional integration. Subgroup analysis by medication status uncovered distinct BNS signatures for medicated and unmedicated FEP. BNS metrics were correlated with social cognition measures in CTR and with positive formal thought disorder in FEP, whereas static metrics showed no such associations. These findings suggest that relative to static connectivity metrics, dynamic connectivity provides non-redundant information and underscores the impact of medication on neural dynamics in psychosis.