ABNORMAL STRUCTURAL COVARIANCE NETWORK IN TREATMENT- RESISTANT DEPRESSION

Abstract Background Treatment-resistant depression (TRD) is associated with gray matter volume reduction compared to healthy controls (HC). Since depression has increasingly been recognized as a disorder of dysregulated neural networks, it is important to evaluate these structural changes in terms of brain networks. Structural covariance is a measurement that gauges the association strength of structural measures between two regions, which is used to examine network-level alteration of brain structure. Although several studies reported abnormal structural covariance in depression, few studies have thus far investigated structural covariance alteration in relation to TRD. Aims and Objectives In this study, we aimed to examine structural covariance alteration in TRD compared to HC. Methods T1 images were obtained from 116 patients with TRD and 54 HC who were recruited at Keio University. Brain volume was calculated for each of 85 regions of the DK atlas using FreeSurfer 6.0. Group differences in brain volumes were examined by analysis of covariance controlling for age and sex. Structural covariances of brain volumes across 3570 pairs of brain regions were obtained by partial correlation analysis using age, sex, and total intracranial volume as covariates. Network-based statistics were used to extract networks with group differences in structural covariances. This study was conducted after obtaining approval from the respective ethics review committees. Results There were no significant differences in brain volumes between the TRD and HC groups. The null hypothesis of equality in structural covariance between them was rejected using network-based statistics. We found a single network comprising connections with elevated structural covariance in patients with TRD compared with HC. The left nucleus accumbens, bilateral pericalcarine, and supramarginal gyrus had high degree centrality in the structural covariance network. Discussion and Conclusion Patients with TRD showed coordinated brain volume alterations in comparison with HC, even though there was no difference in brain volumes between them. Nucleus accumbens, an integral hub in the reward circuit, was one of the hubs of extracted network with elevated structural covariance in TRD, which is in line with a previous study that reported higher structural covariance of nucleus accumbens in depression. Further studies are warranted to examine structural covariance differences among TRD, non-TRD, and HC in order to confirm the link between structural alteration and pathophysiology of treatment resistance in depression.