Abstract / Summary
Resting-state functional connectivity (RSFC) abnormalities are central to characterising brain network dysfunction in first-episode psychosis (FEP), yet biomarker development remains limited by an incomplete understanding of connectivity heterogeneity. This systematic review synthesised 75 studies investigating RSFC in antipsychotic-naïve FEP across five functional network domains and one developmental subgroup. Antipsychotic-naïve FEP was consistently associated with hypoconnectivity within the default mode, salience, frontostriatal, cerebello-thalamo-cortical, and hippocampal networks. Exploratory analyses examined whether global signal regression (GSR) and motion-correction stringency were associated with differences in reported connectivity direction. Motion-control stringency showed no significant association with connectivity direction, indicating that the hypoconnectivity finding was not attributable to motion-related artefact; GSR use was strongly confounded with geographic origin, precluding independent assessment of its association with connectivity direction. Notably, no included study employed within-study preprocessing sensitivity analyses, and cross-national harmonisation efforts remain minimal. Future research should prioritise within-study GSR sensitivity analyses, establish test–retest reliability and external classifier validation for RSFC biomarkers, adopt harmonised preprocessing conventions, and systematically explore sources of directional heterogeneity across preprocessing strata to enable robust cross-national biomarker synthesis.
Keywords: Connectomics; global signal regression; motion artefact; thalamocortical dysconnectivity; salience network; default mode network.