Abstract

The neurobiological heterogeneity of schizophrenia is widely accepted, but it is unclear how mechanistic differences converge to produce the observed phenotype. Establishing a pathophysiological model that accounts for both heterogeneity and phenotypic similarity is essential to inform stratified treatment approaches. In this cross-sectional diffusion tensor imaging (DTI) study, we recruited 77 healthy controls (HC), and 71 patients with DSM-IV diagnosis of schizophrenia (SCZ), and reconstructed the structural connectivity of 90 brain regions covering entire cerebral cortex. We first confirmed the heterogeneity in structural connectivity by showing a reduced inter-individual similarity in SCZ compared with HC. Moreover, we found it was not possible to cluster patients into subgroups with shared patterns of dysconnectivity, indicating a high degree of mechanistic divergence in schizophrenia. Instead of the strength of connectivity between any particular brain regions, we investigated the diversity (or statistically, the variance) of the topographic distribution of the strength was reduced. HC had higher topographic diversity in whole brain structural connectivity compared to the patient group\n\n(P = 2 x 10-6, T = 4.96, Cohen'S d = 0.87). In 62 of the 90 brain regions, the topographic diversity was significantly reduced in patients compared to controls after FDR correction (<0.05). When topographic diversity was used as a discriminant feature for classification between patients and controls, we significantly (P = 4.29 x 10-24) improved the classification accuracy to 79.6% (sensitivity 78.3%, specificity 81.3%). This finding suggests highly individualized pattern of structural dysconnectivity underlying the heterogeneity of schizophrenia converges to a convergent common pathway as reduced topographic diversity for the clinical construct of the disease.