Neurobiological subtypes of adolescent depression: a multimodal integration of morphometric similarity network and spatial transcriptomics
P. Wu, Lingtao Kong, Yifang Zhou, Caijiu Deng, Ziyi Wang, Yuxin Shen, Lei Wang, Zhengjiao Tuo, Yuang Liu, Yucheng Wang, Yuning Zhou, Qikun Sun, Yanqing Tang
Abstract
Adolescent major depressive disorder (AMDD) is a heterogeneous condition with rising global prevalence and limited treatment efficacy. This study integrates morphometric similarity networks (MSN) and spatial transcriptomics to identify neurobiologically distinct AMDD subtypes and their underlying molecular mechanisms. Using the HYDRA algorithm, we delineate two subtypes: AMDD1, characterized by reduced MSN strength in frontoparietal networks, heightened impulsivity, and preserved cognition; and AMDD2, marked by elevated MSN strength in limbic-visual circuits, severe emotional dysregulation, and rumination. Transcriptomic analyses reveal subtype-specific gene expression patterns, with AMDD1 associated with synaptic pruning deficits and AMDD2 linked to GABAergic inhibition deficits. Cell-type mapping highlights astrocytic dysregulation in AMDD1 and microglial activation in AMDD2, while pathway enrichment identifies distinct molecular networks, including endocannabinoid signaling in AMDD1 and MAPK-driven neuroinflammation in AMDD2. Developmental trajectory analysis uncovers critical windows for intervention, with AMDD1 showing delayed cerebellar maturation and AMDD2 exhibiting early hippocampal-striatal priming. These findings advance a precision framework for AMDD, linking spatially patterned gene expression to neurodevelopmental trajectories and offering targeted therapeutic strategies tailored to subtype-specific mechanisms. By bridging molecular, cellular, and network-level insights, this study provides a transformative approach to understanding and treating adolescent depression.