A multiscale characterization of cortical shape asymmetries in early psychosis

Psychosis has often been linked to abnormal cortical asymmetry, but prior results have been inconsistent. Here, we applied a novel spectral shape analysis to characterize cortical shape asymmetries in patients with early psychosis across different spatial scales. We used the Human Connectome Project for Early Psychosis dataset (aged 16-35), comprising 56 healthy controls (37 males, 19 females) and 112 patients with early psychosis (68 males, 44 females). We quantified shape variations of each hemisphere over different spatial frequencies and applied a general linear model to compare differences between healthy controls and patients with early psychosis. We further used canonical correlation analysis to examine associations between shape asymmetries and clinical symptoms. Cortical shape asymmetries, spanning wavelengths from about 22 to 75 mm, were significantly different between healthy controls and patients with early psychosis (Cohen's <i>d</i> = 0.28-0.51), with patients showing greater asymmetry in cortical shape than controls. A single canonical mode linked the asymmetry measures to symptoms (canonical correlation analysis <i>r</i> = 0.45), such that higher cortical asymmetry was correlated with more severe excitement symptoms and less severe emotional distress. Significant group differences in the asymmetries of traditional morphological measures of cortical thickness, surface area, and gyrification, at either global or regional levels, were not identified. Cortical shape asymmetries are more sensitive than other morphological asymmetries in capturing abnormalities in patients with early psychosis. These abnormalities are expressed at coarse spatial scales and are correlated with specific symptom domains.