Department of Neurology, Max Planck Institute for Human Cognitive and Brain Sciences, Leipzig, Saxony, Germany.
Faculty of Computer Science and Media, Leipzig University of Applied Science, Leipzig, Saxony, Germany.
PLoS One. 2020 May 14;15(5):e0228119. doi: 10.1371/journal.pone.0228119. eCollection 2020.
Simulating transcranial electric stimulation is actively researched as knowledge about the distribution of the electrical field is decisive for understanding the variability in the elicited stimulation effect. Several software pipelines comprehensively solve this task in an automated manner for standard use-cases. However, simulations for non-standard applications such as uncommon electrode shapes or the creation of head models from non-optimized T1-weighted imaging data and the inclusion of irregular structures are more difficult to accomplish. We address these limitations and suggest a comprehensive workflow to simulate transcranial electric stimulation based on open-source tools. The workflow covers the head model creation from MRI data, the electrode modeling, the modeling of anisotropic conductivity behavior of the white matter, the numerical simulation and visualization. Skin, skull, air cavities, cerebrospinal fluid, white matter, and gray matter are segmented semi-automatically from T1-weighted MR images. Electrodes of arbitrary number and shape can be modeled. The meshing of the head model is implemented in a way to preserve the feature edges of the electrodes and is free of topological restrictions of the considered structures of the head model. White matter anisotropy can be computed from diffusion-tensor imaging data. Our solver application was verified analytically and by contrasting the tDCS simulation results with that of other simulation pipelines (SimNIBS 3.0, ROAST 3.0). An agreement in both cases underlines the validity of our workflow. Our suggested solutions facilitate investigations of irregular structures in patients (e.g. lesions, implants) or new electrode types. For a coupled use of the described workflow, we provide documentation and disclose the full source code of the developed tools.
经颅电刺激的模拟正在被积极研究，因为电场的分布知识对于理解刺激效果的可变性至关重要。有几个软件管道可以全面地以自动化方式解决这个标准用例任务。然而，对于非标准应用，如不常见的电极形状，或从非优化的 T1 加权成像数据创建头部模型以及包括不规则结构的模拟，就更加困难。我们解决了这些限制，并提出了一个基于开源工具的经颅电刺激模拟的综合工作流程。该工作流程涵盖了从 MRI 数据创建头部模型、电极建模、白质各向异性电导率行为建模、数值模拟和可视化。通过半自动分割 T1 加权 MRI 图像，可以对皮肤、颅骨、气腔、脑脊液、白质和灰质进行分割。可以对任意数量和形状的电极进行建模。头部模型的网格划分方式保留了电极的特征边缘，并且不受头部模型结构的拓扑限制。可以从扩散张量成像数据中计算白质各向异性。我们的求解器应用程序已经通过分析和与其他模拟管道（SimNIBS 3.0、ROAST 3.0）的 tDCS 模拟结果进行对比得到了验证。这两种情况下的一致性都强调了我们工作流程的有效性。我们提出的解决方案为研究患者中的不规则结构（例如病变、植入物）或新型电极类型提供了便利。为了结合使用描述的工作流程，我们提供了文档，并公开了开发工具的完整源代码。
