Resting-State fMRI Graph Theory Analysis for Predicting Selective Serotonin Reuptake Inhibitors Treatment Response in Adolescent Major Depressive Disorder

Xue Mo¹Xuemei Li²Mengqi Liu¹Linlin Hu¹Qian Li¹Jie Wang¹Haiqin Deng¹Fajin Lv¹Xinyu Zhou²Yun Mao^1*Yang Huang^1*

• ¹Department of Radiology, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China
• ²Department of Psychiatry, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China

Background: Substantial interindividual variability exists in the response of adolescents with major depressive disorder (MDD) to selective serotonin reuptake inhibitors (SSRIs), and reliable early predictors of treatment response are lacking. Methods: Resting-state functional magnetic resonance imaging (fMRI) data and clinical scale scores were collected from 69 adolescents with first-episode, drug-naïve MDD. Based on treatment response assessed after 8 weeks of SSRIs therapy, participants were categorized into a responder group (n=37) and a non-responder group (n=32). Graph-theoretical analysis was then performed on the pre-treatment resting-state functional networks of both groups. Results: Significant group differences emerged in several global attribute metrics and multiple brain region node attribute metrics (including the left middle frontal gyrus, hippocampus, parahippocampal gyrus, amygdala, pallidum, as well as the right anterior cingulate cortex and inferior parietal lobule). Partial correlation analyses revealed negative correlations between nodal efficiency in the left middle frontal gyrus, hippocampus, and parahippocampal gyrus, as well as degree centrality in the right anterior cingulate gyrus, and the reduction rate in Hamilton Depression Rating Scale-17 score. Furthermore, logistic regression analysis identified lower nodal efficiency in the right inferior parietal lobule and higher clustering coefficient in the left pallidum as significant predictors of SSRIs treatment response. Conclusions: Pre-treatment functional network topological metrics differentiating responders and non-responders demonstrate potential as predictors for SSRIs treatment response in adolescents with MDD.