Luxian Lv

Higher frequencies of CNV deletions in bipolar disorder among Chinese Population.

Enhanced multimodal MRI classification of schizophrenia through cross-attention graph neural networks.

We propose TB-GCAN, a tri-branch cross-attention graph neural network for schizophrenia classification using multimodal MRI, including sMRI, fMRI, and DTI. Built on a multi-site dataset of 1191 samples from seven scanning sites, the model exploits atlas-defined one-to-one anatomical correspondence across modalities to enable node-level cross-attention during intermediate representation learning. In 7-site leave-one-site-out evaluation, TB-GCAN achieved 84.63% accuracy and outperformed GAT, GCN, CNN, SVM, and MMGNN in the tri-modal setting. Attention-based region ranking highlighted biologically plausible schizophrenia-related regions, and downstream analyses linked the learned imaging representations to PANSS dimensions and transcriptional programs. Unlike generic multimodal GNNs that learn cross-modal relations from data, TB-GCAN directly leverages atlas-aligned regional correspondence to perform anatomically constrained node-level interaction. These findings indicate that anatomically grounded node-level multimodal fusion can improve classification performance while preserving neurobiological interpretability, thereby providing a principled framework for multimodal schizophrenia classification and biomarker discovery.

Integrative transcriptomic and machine learning analysis identifies S100A8 as a key hub bridging peripheral inflammation and synaptic deficits in schizophrenia.

Schizophrenia (SCZ) involves immune dysregulation and synaptic deficits, yet the molecular link between peripheral inflammation and central synaptic pathology remains unclear. We integrated blood transcriptomes from four SCZ cohorts (478 samples: 245 patients, 233 controls) and validated findings in single-cell brain data and an ELNI mouse model. After batch correction, 272 genes were differentially expressed, with S100A8 as the top upregulated immune gene. WGCNA identified a disease-associated module (blue, 133 genes, r = 0.16, p = 7 × 10⁻⁴) containing S100A8; its intersection with differentially expressed genes yielded 44 key genes enriched for cytoplasmic translation, mitochondrial electron transport, and innate immunity. Machine learning ranked S100A8 as the top discriminative feature, and its upregulation was robust across four independent analytic pipelines. CIBERSORTx revealed increased neutrophils and decreased regulatory T and resting NK cells in patients, with S100A8 correlating positively with neutrophils. Single-cell analysis showed that S100A8-expressing cells were specifically expanded in microglia (2.98% → 3.94%, OR = 1.34), and S100A8⁺ microglia displayed an activated state with coordinated upregulation of complement (C1QA/B/C) and phagocytic genes and downregulation of homeostatic markers (P2RY12, CX3CR1). In ELNI mice exhibiting SCZ-like behaviors, S100A8/S100A9 were upregulated in hippocampus and frontal cortex, accompanied by CD68 induction and reduced synaptophysin, with S100A8 correlating positively with CD68 and negatively with synaptophysin. Molecular docking identified hydroxyzine as a candidate S100A8 ligand. These convergent findings establish S100A8 as a hub linking peripheral immune dysregulation to microglial activation and synaptic pathology in SCZ, highlighting it as a candidate biomarker and therapeutic target.