Department of Bioengineering, School of Engineering and Applied Science, University of Pennsylvania, Philadelphia, PA, USA.
Department of Psychology, College of Arts and Sciences, University of Pennsylvania, Philadelphia, PA, USA.
Nat Biomed Eng. 2019 Nov;3(11):902-916. doi: 10.1038/s41551-019-0404-5. Epub 2019 May 27.
Electrocorticography (ECoG) data can be used to estimate brain-wide connectivity patterns. Yet, the invasiveness of ECoG, incomplete cortical coverage, and variability in electrode placement across individuals make the network analysis of ECoG data challenging. Here, we show that the architecture of whole-brain ECoG networks and the factors that shape it can be studied by analysing whole-brain, interregional and band-limited ECoG networks from a large cohort-in this case, of individuals with medication-resistant epilepsy. Using tools from network science, we characterized the basic organization of ECoG networks, including frequency-specific architecture, segregated modules and the dependence of connection weights on interregional Euclidean distance. We then used linear models to explain variabilities in the connection strengths between pairs of brain regions, and to highlight the joint role, in shaping the brain-wide organization of ECoG networks, of communication along white matter pathways, interregional Euclidean distance and correlated gene expression. Moreover, we extended these models to predict out-of-sample, single-subject data. Our predictive models may have future clinical utility; for example, by anticipating the effect of cortical resection on interregional communication.
脑电描记术（ECoG）数据可用于估计大脑范围内的连接模式。然而，ECoG 的侵入性、皮质覆盖的不完整性以及个体之间电极放置的可变性使得 ECoG 数据的网络分析具有挑战性。在这里，我们通过分析来自大样本的全脑、区域间和带限 ECoG 网络，展示了全脑 ECoG 网络的架构以及塑造它的因素。在这种情况下，我们研究了药物难治性癫痫患者的个体。我们使用网络科学的工具来描述 ECoG 网络的基本组织，包括特定频率的架构、隔离的模块以及连接权重对区域间欧几里得距离的依赖性。然后，我们使用线性模型来解释脑区之间连接强度的可变性，并突出沿白质通路的通信、区域间欧几里得距离和相关基因表达在塑造 ECoG 网络的全脑组织中的共同作用。此外，我们将这些模型扩展到预测样本外的单个主体数据。我们的预测模型可能具有未来的临床应用价值；例如，通过预测皮质切除对区域间通信的影响。
