大脑神经振荡动力学机制及其在癫痫疾病中的应用研究

Neural oscillations refer to rhythmic or repetitive neural behavior of discharges in the central nervous system, and it offers an objective description of the electrical activity in the brain.Brain functions are bound up with the generation and modulation of the spatiotemporal features in neural oscillations,while the underlying dynamical mechanism needs further research especially those that related with neurological diseases.According to the preliminary study in the research group, both the frequency component and coupling strength are different between intra and extra focus of brain regions based on cross frequency coupling analysis of intracranial EEG recordings in epilepsy. Taking it as a staring point and targeting at the epilepsy disease, this project will study the frequency dependent information transmission principle as well as the underlying dynamical mechanism of neural oscillations in three aspects divided as neuron and neurotransmitter,information transition of neural firing in circuits and complex network performance of the brain.Proceeding from the internal factors of neurons that affect neural burst firings, we will investigate the influence of neural bursts on frequency response of information transfer in neuron populations, and then we will move forward to the study of interactions between neural network structure and behavior regulation of their vibrational resonance in the network, and all this will contribute to the in-depth knowledge of neural coding and information transition principle related with epileptic seizures, which are accompanied by the frequency dependent feature of the brain network's capability.This project will provide new theory for the understanding of dynamical mechanism of neural oscillations in the brain and offer technical support for the clinical diagnosis assistance, therapeutic effect evaluation and intervention strategy invention.

神经振荡指中枢神经系统神经电活动的重复性或节律活动，是大脑电活动的客观表现。神经振荡时空特性的产生和调节与大脑功能密切相关，但其在神经系统疾病的动力学机制迄今仍需要进一步探讨。课题组前期研究发现在癫痫疾病中，病灶内外脑区在癫痫发作时颅内脑电互频率耦合作用频率成分及作用强度均不同。本项目以此为切入点，选择癫痫疾病为研究对象，从神经元和神经递质、神经回路放电信息传递、大脑复杂网络性能三个层面研究大脑的频率依赖信息传递特征及潜在动力学机制。从神经元阵发放电活动中的神经元内在影响因素出发，分析阵发放电活动对神经元集群信息传递中频率响应的影响规律，并进一步研究神经网络结构与神经元振动共振行为调控的相互作用，以深入了解癫痫发作相关阵发放电活动编码与传递规律及大脑复杂网络性能的频率依赖特征，为揭示大脑神经振荡动力学机制提供新理论，并为癫痫疾病的临床辅助诊断、疗效评估和干预策略开发提供技术支持。

本项目以神经振荡为分析对象，研究癫痫状态下大脑信息传递动力学机制，研究取得了以下成果：1）由于临床癫痫患者脑电信号采集干扰复杂且强度较大，研究了多通道脑电信号预处理的新方法，该方法融合了快速独立成分分析和小波包变换，实现了多通道脑电信号工频干扰去除、小波消噪、分频滤波、奇异值检测等预处理功能；引入相位锁定值和最小生成树算法，实现动态加权频率依赖的复杂网络可视化。2）研究了癫痫脑网络动态连接特征提取及大脑的频率依赖信息传递规律。引入滑动时间窗和希尔伯特变换，利用相位锁定值提取δ、θ、α、β、γ等频段功能连接脑网络，结果发现各波段网络效率、节点强度、网络传递性度特征变化对癫痫发作检测水平差异突出；联合目标攻击和随机误差的弹性检验结果显示，不同类型的癫痫脑网络灵活性存在频率依赖性，频率依赖的信息传递和整合影响大脑功能网络，这可能是大脑信息路由机制之一。3）研究了客观度量大脑功能网络动态变化的新方法。引入资源分配算法对网络节点相似度排序，具有最大相似度的节点形成节点簇, 为观察大脑功能状态变化提供了新途径；在此基础上，引入余弦相似度，提出了一种刻画大脑动态连接的新指标，临床结果显示该指标能有效检测并度量发作过程中大脑网络动力学变化。 4）研究了图像自动拼接和自动分割方法。通过相位相关算法确定粗略重叠区域，利用Harris算法提取特征点和角点邻域灰度信息，探讨全脊柱及下肢X线图像自动拼接算法的可行性，实现了图像的精确配准；引入卷积神经网络，通过归一化互相关技术，运用差异图像、Hough 变换及Ｋ均值聚类，实现了对自由呼吸状态下心脏磁共振成像舒张末期和收缩末期的自动检测。为未来进行大脑多模态技术分析奠定了技术基础。这些成果为进一步揭示癫痫疾病的神经动力学机制提供了理论依据和实践基础。在项目执行期间，已发表论文11篇，其中国外期刊3篇，国内核心学术期刊3篇，国际学术会议分组论文2篇，其中SCI检索3篇，EI检索3篇。