Department of Research and Development, Kempenhaeghe, Sterkselseweg 65, 5591 VE Heeze, The Netherlands.
Neuroimage. 2012 May 1;60(4):2042-53. doi: 10.1016/j.neuroimage.2012.02.023. Epub 2012 Feb 17.
EEG-correlated functional MRI (EEG-fMRI) has been used to indicate brain regions associated with interictal epileptiform discharges (IEDs). This technique enables the delineation of the complete epileptiform network, including multifocal and deeply situated cortical areas. Before EEG-fMRI can be used as an additional diagnostic tool in the preoperative work-up, its added value should be assessed in relation to intracranial EEG recorded from depth electrodes (SEEG) or from the cortex (ECoG), currently the clinical standard. In this study, we propose a framework for the analysis of the SEEG data to investigate in a quantitative way whether EEG-fMRI reflects the same cortical areas as identified by the IEDs present in SEEG recordings. For that purpose, the data of both modalities were analyzed with a general linear model at the same time scale and within the same spatial domain. The IEDs were used as predictors in the model, yielding for EEG-fMRI the brain voxels that were related to the IEDs and, similarly for SEEG, the electrodes that were involved. Finally, the results of the regression analysis were projected on the anatomical MRI of the patients. To explore the usefulness of this quantitative approach, a sample of five patients was studied who both underwent EEG-fMRI and SEEG recordings. For clinical validation, the results of the SEEG analysis were compared to the standard visual review of IEDs in SEEG and to the identified seizure onset zone, the resected area, and outcome of surgery. SEEG analysis revealed a spatial pattern for the most frequent and dominant IEDs present in the data of all patients. The electrodes with the highest correlation values were in good concordance with the electrodes that showed maximal amplitude during those events in the SEEG recordings. These results indicate that the analysis of SEEG data at the time scale of EEG-fMRI, using the same type of regression model, is a promising way to validate EEG-fMRI data. In fact, the BOLD areas with a positive hemodynamic response function were closely related to the spatial pattern of IEDs in the SEEG recordings in four of the five patients. The areas of significant BOLD that were not located in the vicinity of depth electrodes, were mainly characterized by negative hemodynamic responses. Furthermore, the area with a positive hemodynamic response function overlapped with the resected area in three patients, while it was located at the edge of the resection area for one. To conclude, the results of this study encourage the application of EEG-fMRI to guide the implantation of depth electrodes as prerequisite for successful epilepsy surgery.
脑电相关功能磁共振成像 (EEG-fMRI) 已被用于指示与癫痫样放电 (IEDs) 相关的脑区。该技术能够描绘完整的癫痫样网络，包括多灶性和深部皮质区域。在 EEG-fMRI 可以作为术前评估的附加诊断工具之前，应该评估其相对于颅内电极记录 (SEEG) 或皮质记录 (ECoG) 的附加值，目前这是临床标准。在这项研究中，我们提出了一种用于分析 SEEG 数据的框架，以便以定量方式研究 EEG-fMRI 是否反映了与 SEEG 记录中存在的 IED 相同的皮质区域。为此，在相同的时间尺度和相同的空间域内，同时使用广义线性模型分析两种模态的数据。将 IED 用作模型中的预测因子，从而为 EEG-fMRI 生成与 IED 相关的脑体素，并且对于 SEEG 类似，生成涉及的电极。最后，将回归分析的结果投影到患者的解剖 MRI 上。为了探索这种定量方法的有用性，研究了五个同时接受 EEG-fMRI 和 SEEG 记录的患者的样本。为了进行临床验证，将 SEEG 分析的结果与 SEEG 中 IED 的标准视觉评估以及确定的发作起始区、切除区域和手术结果进行比较。SEEG 分析揭示了所有患者数据中最常见和最主要的 IED 的空间模式。具有最高相关值的电极与 SEEG 记录中这些事件期间显示最大振幅的电极高度一致。这些结果表明，在 EEG-fMRI 的时间尺度上使用相同类型的回归模型分析 SEEG 数据是验证 EEG-fMRI 数据的一种很有前途的方法。实际上，在五个患者中的四个中，具有正血流动力学响应功能的 BOLD 区域与 SEEG 记录中 IED 的空间模式密切相关。不在深部电极附近的显著 BOLD 区域主要具有负血流动力学响应特征。此外，具有正血流动力学响应功能的区域与三个患者的切除区域重叠，而一个患者的该区域位于切除区域的边缘。总之，这项研究的结果鼓励将 EEG-fMRI 应用于指导深部电极的植入，作为成功进行癫痫手术的前提。
