Study on the changes of Structural Covariance Network in post-traumatic stress disorder.

Abstract

The topological properties of functional brain networks in post-traumatic stress disorder (PTSD) have been thoroughly examined, whereas the topology of structural covariance networks has been researched much less. Based on graph theoretical approaches, we investigated the topological architecture of structural covariance networks among PTSD, trauma-exposed controls (TEC), and healthy controls (HC) by constructing covariance networks driven by inter-regional correlations of cortical thickness. Structural magnetic resonance imaging (sMRI) scans and clinical scales were performed on 27 PTSD, 33 TEC, and 29 HC subjects. Group-level structural covariance networks were established using pearson correlations of cortical thickness between 68 brain areas, and the graph theory method was utilized to study the global and nodal properties. PTSD and HC subjects did not differ at the global level. When PTSD subjects were compared to TEC subjects, they had significantly higher clustering coefficient (p = .014) and local efficiency (p = .031). Nodes having different nodal centralities between groups did not pass the false-discovery rate correction at the node level. According to the structural brain network topological characteristics discovered in this study, PTSD manifests differently compared to the TEC group. In the PTSD group, the SCN keeps the small-world characteristics, but the degree of functional separation is enhanced. The TEC group's reduced small worldness and the tendency for brain network randomization could be signs of trauma recovery.