A study on the potential relationship between the diagnosis and functional connectivity in the brain in major depressive disorder

Many studies on resting-state functional connectivity (FC) in major depressive disorder (MDD) have investigated FC as a biomarker for disease pathogenesis. However, few studies have examined conditional dependencies among FC, clinical status, and demographic variables. Using a large dataset comprising 431 healthy controls and 235 MDD patients, we combined partial least squares (PLS)-based feature extraction with logistic regression and LightGBM models for the diagnosis, followed by Bayesian network (BN) analysis to explore underlying structures using a directed acyclic graph. The classification models achieved moderate performance (logistic regression: Area Under the Curve (AUC) of 0.735; LightGBM: AUC of 0.710). Structure learning using the Max–Min Hill-Climbing algorithm revealed direct edges from the MDD diagnosis to variables extracted based on BDI and PLS, but no variables were identified as direct parent nodes of MDD. Intervention simulation showed that the MDD diagnosis significantly reduced functional connectivity (FC) in the default mode network (DMN), dorsal attention network, and between subcortical structures and cortex by 6.956. These findings suggest that the MDD diagnosis is strongly associated with disease-specific alterations not only in the DMN but also in various networks, emphasizing the need to consider widespread network dysfunction in understanding the pathophysiology of MDD. However, as this study demonstrated statistical associations based on BNs, future research using longitudinal data and intervention studies is needed to more definitively elucidate the relationship between the diagnosis and brain function, as well as their mutual influence.