Noirhomme Quentin, Ferrant Matthieu, Vandermeeren Yves, Olivier Etienne, Macq Benoît, Cuisenaire Olivier
Communications and Remote Sensing Laboratory, Université catholique de Louvain, 1348 Louvain-la-Neuve, Belgium.
IEEE Trans Biomed Eng. 2004 Nov;51(11):1994-2005. doi: 10.1109/TBME.2004.834266.
This paper describes a method for registering and visualizing in real-time the results of transcranial magnetic stimulations (TMS) in physical space on the corresponding anatomical locations in MR images of the brain. The method proceeds in three main steps. First, the patient scalp is digitized in physical space with a magnetic-field digitizer, following a specific digitization pattern. Second, a registration process minimizes the mean square distance between those points and a segmented scalp surface extracted from the magnetic resonance image. Following this registration, the physician can follow the change in coil position in real-time through the visualization interface and adjust the coil position to the desired anatomical location. Third, amplitude of motor evoked potentials can be projected onto the segmented brain in order to create functional brain maps. The registration has subpixel accuracy in a study with simulated data, while we obtain a point to surface root-mean-square error of 1.17+/-0.38 mm in a 24 subject study.
本文描述了一种用于在物理空间中对经颅磁刺激(TMS)结果进行实时配准并在大脑磁共振图像的相应解剖位置上进行可视化的方法。该方法主要包括三个步骤。首先,使用磁场数字化仪按照特定的数字化模式在物理空间中对患者头皮进行数字化。其次,配准过程使这些点与从磁共振图像中提取的分割头皮表面之间的均方距离最小化。完成此配准后,医生可以通过可视化界面实时跟踪线圈位置的变化,并将线圈位置调整到所需的解剖位置。第三,可以将运动诱发电位的幅度投影到分割后的大脑上,以创建功能性脑图谱。在一项模拟数据研究中,配准具有亚像素精度,而在一项针对24名受试者的研究中,我们获得的点到表面的均方根误差为1.17±0.38毫米。