Depression is a prevalent mental disorder characterized by persistent negative mood and loss of pleasure. Although there are various treatment modalities available for depression, the rates of response and remission remain low. Xiaoyaosan (XYS), a traditional Chinese herbal formula with a long history of use in treating depression, has shown promising effects. However, the underlying mechanism of its therapeutic action remains elusive. The aim of this study is to investigate the neuroimaging changes in the brain associated with the antidepressant-like effects of XYS.
Here, we combined voxel-based morphometry of T2-weighted images and voxel-based analysis on diffusion tensor images to evaluate alterations in brain morphometry and microstructure between chronic social defeat stress (CSDS) model mice and control mice. Additionally, we examined the effect of XYS treatment on structural disruptions in the brains of XYS-treated mice. Furthermore, we explored the therapeutic effect of 18β-glycyrrhetinic acid (18β-GA), which was identified as the primary compound present in the brain following administration of XYS. Significant differences in brain structure were utilized as classification features for distinguishing mice with depression model form the controls using a machine learning method.
Significant changes in brain volume and diffusion metrics were observed in the CSDS model mice, primarily concentrated in the nucleus accumbens (ACB), primary somatosensory area (SSP), thalamus (TH), hypothalamus (HY), basomedical amygdala nucleus (BMA), caudoputamen (CP), and retrosplenial area (RSP). However, both XYS and 18β-GA treatment prevented disruptions in brain volume and diffusion metrics in certain regions, including bilateral HY, right SSP, right ACB, bilateral CP, and left TH. The classification models based on each type of neuroimaging feature achieved high accuracy levels (gray matter volume: 76.39%, AUC=0.83; white matter volume: 76.39%, AUC=0.92; fractional anisotropy: 82.64%, AUC=0.9; radial diffusivity: 76.39%, AUC=0.82). Among these machine learning analyses, the right ACB, right HY, and right CP were identified as the most important brain regions for classification purposes.
These findings suggested that XYS can prevent abnormal changes in brain volume and microstructure within TH, SSP, ACB, and CP to exert prophylactic antidepressant-like effects in CSDS model mice. The neuroimaging features within these regions demonstrate excellent performance for classifying CSDS model mice from controls while providing valuable insights into the antidepressant effects of XYS.