Exploring Schizophrenia Classification Through Multimodal MRI and Deep Graph Neural Networks: Unveiling Brain Region-Specific Weight Discrepancies and Their Association With Cell-Type Specific Transcriptomic Features.

BACKGROUND AND HYPOTHESIS: Schizophrenia (SZ) is a prevalent mental disorder that imposes significant health burdens. Diagnostic accuracy remains challenging due to clinical subjectivity. To address this issue, we explore magnetic resonance imaging (MRI) as a tool to enhance SZ diagnosis and provide objective references and biomarkers. Using deep learning with graph convolution, we represent MRI data as graphs, aligning with brain structure, and improving feature extraction, and classification. Integration of multiple modalities is expected to enhance classification. STUDY DESIGN: Our study enrolled 683 SZ patients and 606 healthy controls from 7 hospitals, collecting structural MRI and functional MRI data. Both data types were represented as graphs, processed by 2 graph attention networks, and fused for classification. Grad-CAM with graph convolution ensured interpretability, and partial least squares analyzed gene expression in brain regions. STUDY RESULTS: Our method excelled in the classification task, achieving 83.32% accuracy, 83.41% sensitivity, and 83.20% specificity in 10-fold cross-validation, surpassing traditional methods. And our multimodal approach outperformed unimodal methods. Grad-CAM identified potential brain biomarkers consistent with gene analysis and prior research. CONCLUSIONS: Our study demonstrates the effectiveness of deep learning with graph attention networks, surpassing previous SZ diagnostic methods. Multimodal MRI's superiority over unimodal MRI confirms our initial hypothesis. Identifying potential brain biomarkers alongside gene biomarkers holds promise for advancing objective SZ diagnosis and research in SZ.

[Changes of brain network characteristics in patients with depression before and after precise repetitive transcranial magnetic stimulation].

Objective: To investigate the changes of brain network characteristics in patients with depression before and after precise repetitive transcranial magnetic stimulation (rTMS) treatment. Methods: Patients with depression in the Second Affiliated Hospital of Xinxiang Medical University and healthy volunteers in the community of Xinxiang city from February 2018 to March 2019 were simultaneously recruited. The left dorsolateral prefrontal cortex was precisely selected as the stimulation target through the latest Human Brainnetome Atlas, and the near infrared navigation was used to achieve accurate brain stimulation treatment in combination with the structural magnetic resonance data. Moreover, functional connectivity was analyzed before and after rTMS treatment in significantly altered brain areas of patients with depression. Results: Nineteen patients (11 males and 8 females) with depression were included, aged (34±11) years. Meanwhile, 22 healthy controls (9 males and 13 females), aged (30±9) years, were also enrolled. Functional connectivity of insular cortex was decreased in depression patients when the insula was analyzed as the target area (P<0.05). The functional connection from insula to middle frontal lobe and superior parietal lobe in patients with depression decreased before rTMS treatment (P<0.05), but increased after rTMS treatment (P<0.05). The functional connection between dIg_L of the insula and the right middle prefrontal lobe was correlated with Beck Anxiety Index (BAI) before rTMS treatment and Beck Depression Index (BDI) after rTMS treatment (r=0.737, P=0.003; r=0.696, P=0.005). Conclusions: Abnormal functional connectivity of insula may be the brain imaging mechanism of rTMS treatment. Precise brain region selection based on Human Brainnetome Atlas provides a new technical method for clinical rTMS precision treatment.

Carbon-Based Fiber Materials as Implantable Depth Neural Electrodes.

Implantable brain electrophysiology electrodes are valuable tools in both fundamental and applied neuroscience due to their ability to record neural activity with high spatiotemporal resolution from shallow and deep brain regions. Their use has been hindered, however, by the challenges in achieving chronically stable operations. Furthermore, implantable depth neural electrodes can only carry out limited data sampling within predefined anatomical regions, making it challenging to perform large-area brain mapping. Minimizing inflammatory responses and associated gliosis formation, and improving the durability and stability of the electrode insulation layers are critical to achieve long-term stable neural recording and stimulation. Combining electrophysiological measurements with simultaneous whole-brain imaging techniques, such as magnetic resonance imaging (MRI), provides a useful solution to alleviate the challenge in scalability of implantable depth electrodes. In recent years, various carbon-based materials have been used to fabricate flexible neural depth electrodes with reduced inflammatory responses and MRI-compatible electrodes, which allows structural and functional MRI mapping of the whole brain without obstructing any brain regions around the electrodes. Here, we conducted a systematic comparative evaluation on the electrochemical properties, mechanical properties, and MRI compatibility of different kinds of carbon-based fiber materials, including carbon nanotube fibers, graphene fibers, and carbon fibers. We also developed a strategy to improve the stability of the electrode insulation without sacrificing the flexibility of the implantable depth electrodes by sandwiching an inorganic barrier layer inside the polymer insulation film. These studies provide us with important insights into choosing the most suitable materials for next-generation implantable depth electrodes with unique capabilities for applications in both fundamental and translational neuroscience research.