Abnormalities in structural covariance of cortical gyrification in schizophrenia

Abstract

The highly convoluted shape of the adult human brain results from several well-coordinated maturational events that start from embryonic development and extend through the adult life span. Disturbances in these maturational events can result in various neurological and psychiatric disorders, resulting in abnormal patterns of morphological relationship among cortical structures (structural covariance). Structural covariance can be studied using graph theory-based approaches that evaluate topological properties of brain networks. Covariance-based graph metrics allow cross-sectional study of coordinated maturational relationship among brain regions. Disrupted gyrification of focal brain regions is a consistent feature of schizophrenia. However, it is unclear if these localized disturbances result from a failure of coordinated development of brain regions in schizophrenia. We studied the structural covariance of gyrification in a sample of 41 patients with schizophrenia and 40 healthy controls by constructing gyrification-based networks using a 3-dimensional index. We found that several key regions including anterior insula and dorsolateral prefrontal cortex show increased segregation in schizophrenia, alongside reduced segregation in somato-sensory and occipital regions. Patients also showed a lack of prominence of the distributed covariance (hubness) of cingulate cortex. The abnormal segregated folding pattern in the right peri-sylvian regions (insula and fronto-temporal cortex) was associated with greater severity of illness. The study of structural covariance in cortical folding supports the presence of subtle deviation in the coordinated development of cortical convolutions in schizophrenia. The heterogeneity in the severity of schizophrenia could be explained in part by aberrant trajectories of neurodevelopment.

Fig. 1

Steps in processing the gyrification-based networks. a Surface reconstruction was carried out using FreeSurfer software; local gyrification indices were estimated using Schaer’s procedure. b Destrieux atlas was used for parcellating the cortical surface to 148 regions (74 on each hemisphere). c Association matrices were obtained by calculating the correlations between regional gyrification across subjects within each group separately. d Binary adjacency matrices were derived from thresholding at minimum density for fully connected graphs in both groups

Fig. 2

Regional changes in topological properties of the gyrification network. A colour figure is provided online. Middle frontal and short insula show increased segregation in patients with schizophrenia (increased in local efficiency/clustering coefficient). Inferior temporal, intraparietal and posterior midcingulate show decreased degree centrality in schizophrenia. All other labelled regions show reduced segregation in patients with schizophrenia. L left hemisphere, R right hemisphere, G gyrus, S sulcus

Fig. 3

Graphical representation of gyrification networks in controls (CON) and patients with schizophrenia (SCZ), visualized using BrainNet viewer (http://www.nitrc.org/projects/bnv). Both CON and SCZ networks had 6 modules each discovered using Newman’s module detection algorithm, coded separately for each network. The size of the nodes is proportional to the degree centrality. A colour figure showing module membership of individual nodes is provided online