Identification of first-episode unmedicated major depressive disorder using pretreatment features of dominant coactivation patterns

Identifying neuroimaging features to diagnose major depressive disorder (MDD) and predict treatment response remains challenging. Using the pretreatment dominant coactivation pattern (dCAP) analysis approach, we aimed to identify patients with MDD and predict antidepressant efficacy. Seventy-seven first-episode unmedicated MDD patients and forty-two age- and sex-matched healthy controls (HCs) were recruited in the study. The dCAP analysis was performed for the reward and default mode network (DMN) to identify the MDD patients from the HCs. The dCAP1 of the left posterior DMN and bilateral anterior DMN were significantly higher in the MDD group than in the HC group (P < .001), and the dCAP1 in the left posterior DMN was positively correlated with the baseline severity of depression (rho = 0.248, P = .030). Besides, the MDD group exhibited significantly higher dCAP1 in the right reward network than the HC group. Further correlation analyses revealed that the transfer probability in the right reward network was positively correlated with the treatment responsivity (r = 0.247, P = .030). Importantly, integrating the dCAPs of the above four subnetworks can effectively identify the patients with MDD (AUC = 0.920, P < .001). The distinct pretreatment features of the dCAP in the subnetwork of the DMN and reward network may serve as potential indicators for individual diagnosis and prediction of antidepressant response in the early stage.

Automated summary

Using the dCAP analysis approach, distinct pretreatment features of the DMN and reward network can effectively differentiate MDD patients from healthy controls, potentially serving as indicators for individual diagnosis and prediction of antidepressant response in the early stage.