"First-episode psychosis: Structural covariance deficits in salience network correlate with symptoms severity".

BACKGROUND:Patterns of coordinated variations of gray matter (GM) morphology across individuals are promising indicators of disease. However, it remains unclear if they can help characterize first-episode psychosis (FEP) and symptoms' severity. METHODS:Sixty-seven FEP and 67 matched healthy controls (HC) were assessed with structural MRI to evaluate the existence of distributed GM structural covariance patterns associated to brain areas belonging to salience network. Voxel-based morphometry (VBM) and structural covariance differences, investigated with salience network seed-based Partial Least Square, were applied to explore differences between groups. GM density associations with Raven's intelligent quotient (IQ) and Positive and Negative Syndrome Scale (PANSS) scores were investigated. RESULTS:Univariate VBM results gave trend without significant GM differences across groups. GM and IQ correlated positively in both groups: in FEP, mostly in hippocampus, insula, and fronto-temporal structures, while in HC mostly in amygdala, thalamus and fronto-temporal regions. GM and PANSS scores correlated negatively in FEP, with widespread clusters located in limbic regions. Multivariate analysis showed strong and opposite structural GM covariance with salience network for FEP and HC. Moreover, structural covariance of the salience network in FEP correlated negatively with severity of clinical symptoms. CONCLUSION:Our study provides evidence supporting the insular dysfunction model of psychosis. Reduced structural GM covariance of the salience network, with its association to symptom's severity, appears a promising morphometry feature for FEP detection.