Chu Na, Wang Dixin, Qu Shanshan, Yan Chang, Luo Gang, Liu Xuesong, Hu Xiping, Zhu Jing, Li Xiaowei, Sun Shuting, Hu Bin
Key Laboratory of Brain Health Intelligent Evaluation and Intervention, Beijing Institute of Technology, Ministry of Education, Beijing 100081, China.
Gansu Provincial Key Laboratory of Wearable Computing, School of Information Science and Engineering, Lanzhou University, Lanzhou 730000, China.
Prog Neuropsychopharmacol Biol Psychiatry. 2025 Jan 10;136:111149. doi: 10.1016/j.pnpbp.2024.111149. Epub 2024 Sep 19.
The modular structure can reflect the activity pattern of the brain, and exploring it may help us understand the pathogenesis of major depressive disorder (MDD). However, little is known about how to build a stable modular structure in MDD patients and how modules are separated and integrated.
We used four independent resting state Electroencephalography (EEG) datasets. Different coupling methods, window lengths, and optimized community detection algorithms were used to find a reliable and robust modular structure, and the module differences of MDD were analyzed from the perspectives of global module attributes and local topology in multiple frequency bands.
The combination of the Phase Lag Index (PLI) and the Louvain algorithm can achieve better results and can achieve stability at smaller window lengths. Compared with Healthy Controls (HC), MDD had higher Modularity (Q) values and the number of modules in low-frequency bands. In addition, MDD showed significant structural changes in the frontal and parietal-occipital lobes, which were confirmed by further correlation analysis.
Our results provided a reliable validation of the modular structure construction method in MDD patients and contributed strong evidence for the changes in emotional cognition and visual system function in MDD patients from a new perspective. These results would afford valuable insights for further exploration of the pathogenesis of MDD.
模块化结构能够反映大脑的活动模式，对其进行探索可能有助于我们理解重度抑郁症（MDD）的发病机制。然而，关于如何在MDD患者中构建稳定的模块化结构以及模块如何分离和整合，我们知之甚少。
我们使用了四个独立的静息态脑电图（EEG）数据集。采用不同的耦合方法、窗口长度和优化的社区检测算法来寻找可靠且稳健的模块化结构，并从多个频段的全局模块属性和局部拓扑结构的角度分析MDD的模块差异。
相位滞后指数（PLI）与Louvain算法相结合能够取得更好的结果，并且在较小的窗口长度下就能实现稳定性。与健康对照（HC）相比，MDD的模块化（Q）值和低频段的模块数量更高。此外，MDD在额叶和顶枕叶显示出显著的结构变化，这通过进一步的相关性分析得到了证实。
我们的结果为MDD患者模块化结构构建方法提供了可靠的验证，并从新的角度为MDD患者情绪认知和视觉系统功能的变化提供了有力证据。这些结果将为进一步探索MDD的发病机制提供有价值的见解。
