毕业生知识库

脑机接口（Brain-computer Interface, BCI）是指在大脑与外部设备之间建立不 依赖于外周神经和肌肉的信息通道，可通过检测（读）或调控（写）神经元活动， 实现大脑与外部设备的信息交互。针对柔性电极植入手术创伤较大的问题，本文 重点研究颅脑影像空间与物理空间的无创配准、激光器光轴标定与开孔路径空间 转换、颅脑规划开孔路径与激光光轴对准导航等关键技术，形成集“无创配准-光 轴标定-空间转换-对准导航”功能于一体的激光微创开孔导航系统。本文的主要 研究工作与贡献如下： （1）针对柔性电极植入手术需植入骨钉、存在出血和感染风险的问题，提 出免植入钉的颅脑影像空间和物理空间无创配准方法，在猕猴颅脑开孔导航应用 场景评测配准误差和定位误差，提出基于混合遗传算法的目标靶点最优配准点集 自动选取方法，有效提升了颅脑空间配准精度。 （2）针对规划开孔路径和开孔激光光轴测量难题，提出激光器光轴空间标 定与拟合方法，聚焦空间转换具体过程，采用空间直线拟合的方法标定激光器光 轴，设计模拟开孔装置用于直观显示颅脑规划开孔路径的位置和姿态，便于后续 实验激光器光轴和规划开孔路径对准开孔操作。最后建立激光器坐标系、颅脑坐 标系和相机坐标系之间的空间转换关系，并通过实验验证了激光器光轴标定和空 间转换的精度，实现颅脑规划开孔路径和激光器光轴位姿的实时定位。 （3）提出基于光电跟踪仪的颅脑开孔路径与激光器光轴位姿对准方法，建 立开孔导航设备位姿测量单元和运动单元的映射关系，设计基于位置的伺服控制 系统引导颅脑规划开孔路径和激光光轴的快速接近、精确对准。实验结果表明： 颅脑规划路径和激光器光轴姿态对准误差≤0.16°，位置对准误差≤0.4mm，实 现了面向植入式脑机接口应用的激光微创开孔自动导航。

The term Brain-computer Interface (BCI) refers to the establishment of an information channel between the brain and external devices that is independent of peripheral nerves and muscles. It enables the exchange of information between the brain and external devices through the detection (reading) or regulation (writing) of neuronal activity. In response to the significant trauma associated with flexible electrode implantation surgery, this paper focuses on key technologies including non-invasive registration of cranial imaging space to physical space, calibration of laser axis and space transformation of drilling pathway, alignment navigation between cranial planning of drilling pathways and laser axis. These efforts culminate in the development of a laser-guided minimally invasive drilling navigation system integrating functions of "non-invasive registration - calibration of laser axis - space transformation - alignment navigation". The main research work and contributions of this paper are as follows: (1) Addressing the issues of implantation surgery of flexible electrodes requiring the implantation of bone nails and posing risks of bleeding and infection, we proposed a non-invasive registration method for cranial imaging space to physical space without the need for implantable nails. The registration error and positioning error were evaluated in the application scenarios of rhesus monkey brain drilling navigation, a method based on a hybrid genetic algorithm for selecting optimal registration marker spheres is presented, thus enhancing registration accuracy between cranial imaging space and physical space. (2) To address the challenges of planning borehole paths and measuring the laser axis for borehole drilling, a laser axis spatial calibration and fitting method is proposed, focusing on the specific process of space conversion, using the spatial linear fitting method to calibrate the laser optical axis, and designing a simulated opening device for visually displaying the position and attitude of the brain planning the drilling path. It is convenient for the subsequent experimental laser optical axis and planning the drilling path to align the opening operation. Finally, the spatial transformation relationship between laser coordinate system, craniocerebral coordinate system and camera coordinate system is established, and the accuracy of laser optical axis calibration and spatial transformation is verified by experiments, and real-time positioning of craniocerebral planning drilling path and laser optical axis pose is realized. (3) This paper proposes a method for aligning the cranial drilling pathway with the laser axis based on a photoelectric tracker. This method establishes a mapping relationship between the pose measurement unit and the motion unit of the drilling navigation device. We design a position-based servo control system to guide the rapid approach and precise alignment of the planned cranial drilling pathway and the laser axis. Experimental results indicate that the alignment error between the planned cranial path and the laser axis attitude is ≤0.16°, and the positional alignment error is ≤ 0.4mm, achieving automated navigation for minimally invasive laser borehole drilling tailored for implantable brain-computer interface applications.