多通道关联脑电信号复合处理与硅芯片集成技术研究

项目名称： 多通道关联脑电信号复合处理与硅芯片集成技术研究

项目编号： No.61271046

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 无线电电子学、电信技术

项目作者： 李洪革

作者单位： 北京航空航天大学

项目金额： 80万元

中文摘要： 脑机信息交互植入式微系统是探索人-机融合的最前沿的课题。近年来，随着微电子技术的发展，植入式无线集成微系统接口成为研究者探索的热点问题。脑电神经信号由于其复杂性、互扰性、时空性和微弱等特性，同时，传统拖缆脑电获取系统不利于实现残疾假肢有效功能代偿等任务，而无线植入式接口在现设计原理和通信带宽限制下，获取脑信号存在严重丢失而难以准确重现原始脑电信号。因此，本课题提出脑电信号的动态非线性能量-复杂布尔网络检测机制及集成化方案。课题以基于非线性能量算子和布尔网络的锋电位检测/分类为突破点，运用信息学、微电子学和神经工程学，重点探索布尔神经网络的脑电信号智能分类及亚阈值超低功耗CMOS集成化建模。在获取多通道空间并行神经元锋电位时，通过阵列动态能量算子检测在无尺度学习响应模式下，实现适应分类的低数据传输率的微弱脑电信号重现的实时性、准确性、复用性，以探讨人-机融合接口系统的可行性。

中文关键词： 脑电处理；多通道；微纳系统芯片；抗EMI；

英文摘要： Microsystem of brain neural recording is advance research about Implantable Brain-Machines Interface. Today, the wireless implantable microsystems become studying challenges since the developing of microelectronic. The conventions neural recording with wire electrodes cannot implement the long-term monitoring application, on the other hand, there is a critical technical challenge to the implementation of wireless implantable recording microsystem for transmitting bandwidth limitation. In order to address this constraint problem before data transmission, the low complexity and high efficient spike sorting method is necessary. Therefore, the authors propose a novel method of brain neural recording with high-accurate, high-robust and low-data rate. The spike detecting and sorting of brain neural signal based on nonlinear energy operator and Boolean network is the studying point, using information processing and microelectronic technology. Its core task is of exploring the multi-overlap signal detection method and microsystem integrated design with Nonlinear Energy Operator-Boolean network. The integration of a significant portion or all of an implantable BMI microsystem onto an implantable platform is a highly desirable for future portable and wearable prosthetics.

英文关键词： brain neural processing；mulltichannel；micro/nano chip；EMI immunity；