Brain network predictors of changes in symptoms and serum BDNF following antidepressant treatment with escitalopram and Yueju Pill in major depressive disorder: a randomised, double-blind, placebo-controlled pilot study

Background Yueju Pill, a classic traditional Chinese medicine, shows antidepressant effects rapidly. However, biomarkers that can predict its treatment outcomes in major depressive disorder (MDD) are still lacking. Multimodal magnetic resonance imaging (MRI) offers a promising avenue to identify such biomarkers.Aims This pilot study aimed to explore whether therapeutic responses to Yueju Pill could be predicted by MRI-derived brain networks and to identify drug-specific biomarkers in comparison to escitalopram, a mainstream antidepressant.Methods We collected multimodal MRI data and blood samples from 28 outpatients with MDD from the Fourth People’s Hospital of Taizhou, who were randomly divided into two groups to receive either Yueju Pill (23 g/time/day) or escitalopram (10 mg, two times a day) for 4 days. Morphological and functional brain networks were constructed and used to predict individual changes in symptoms quantified by the 24-item Hamilton Depression Scale (HAMD-24) scores and serum brain-derived neurotrophic factor (BDNF) levels.Results After the treatment, both groups exhibited significant reductions in the HAMD-24 scores, while only the Yueju Pill group showed significant increases in the BDNF levels. Gyrification Index-based morphological networks predicted change rates of the HAMD-24 scores in both groups, but sulcus depth-based and cortical thickness-based morphological networks predicted change rates of the HAMD-24 scores and BDNF levels, respectively, only in the Yueju Pill group. Subnetwork analyses revealed that the visual network independently predicted the changes in both the HAMD-24 scores (sulcus depth-based networks) and BDNF levels (cortical thickness-based networks) following Yueju Pill treatment.Conclusions Morphological but not functional brain networks can predict symptom improvement and BDNF changes of patients with MDD after Yueju Pill treatment. Sulcus depth-based and cortical thickness-based morphological brain networks, particularly their visual subnetworks, might serve as Yueju Pill-specific biomarkers for predicting the therapeutic responses. These findings have the potential to guide personalised therapy for patients with MDD early in the therapeutic process.Trial registration number ChiCTR1900021114.