经颅磁刺激诱导的脑节律挖掘和分析

Trancranial magnetic stimulation (TMS ) is becoming a novel physical regulation for neurological psychiatric diseases, the efficiency of TMS regulation is evaluated by intergating with EEG (TMS-EEG). Aiming at the key scientific issue of "signal extraction, brain-region localization and functional analysis" in TMS-EEG, this project focuses on the deep mining of brain rhythms with the TMS-induced EEG. This project deeply explores the information of EEG rhythms based on the neural oscillation and synchronization theory. An adaptive decomposition and detection method is first proposed to accurately estimate the instantaneous phase, frequency and magnitude of brain rhythms in real time. Then, a new phase synchronization method is developed to reliably calculate the phase relationship among the EEG rhythms, which is robust to the effect of volume conduction, reference electrode, sample size and noise in EEG.Furthermore, with the dynamical characteristics analysis of brin rhythms, a recognition technique based on a mean field model is proposed to qualitatively analyze the dynamics of deep brain tissue states. Finally, with a Kalman filter, a time-varying brain source estimation method is developed to real-time estimate the brain sources, so as to meet the needs of TMS-EEG. The development of this project will help to expand the applications of TMS, and especially to provide the key technical basis for the applications in areas of brain functional localization and rehabilitation.

经颅磁刺激系统(TMS)成为神经精神疾病的主要物理调控方法, 尤其与脑电采集,可以实时地评价TMS调控效果。针对TMS诱导的脑信号存在的"信号提取，脑区定位和功能分析"核心科学问题，集成脑节律分析方法和计算神经模型，深入挖掘TMS诱导的脑节律信息和功能源。主要包括：基于神经振荡同步理论，深入挖掘TMS诱导的脑电节律信息；提出一自适应分解和检测方法，实时精确估计TMS诱导的脑节律的瞬时相位、频率和幅值；为避免容积传导、电极参考、噪声和样本数量的影响，发展新的计算方法可靠地估计TMS诱导的脑电节律间的耦合关系；发展一基于平均区域模型的深度脑组织状态识别技术，通过脑节律动力学特性分析，定性推测深度TMS诱导的脑组织动力学模式；发展一卡尔曼滤波的时变脑源估计方法，实时估计TMS诱导的脑源信号，以满足TMS定位需要。该课题目标是为扩展TMS应用于脑功能定位和康复等领域应用提供技术支撑。

经颅磁刺激(TMS)因为其无痛无创的特点已经在神经调控领域得到广泛的发展。经颅磁刺激与脑电（EEG）的集成将干预与成像技术融合在一起为探索大脑功能提供了有力的工具。但是二者技术的融合以及TMS下EEG特征的提取与分析仍需要深入探索。因此，本项目主要研究了TMS下EEG信号的获取、处理与分析。并将该技术运用到临床研究中。在本项目结题时，已经形成了一整套完善的TMS下EEG信号的提取处理技术，提出了自己的TMS下EEG去噪算法，从噪声中提取到大脑对TMS刺激响应的真实信号。基于正常人的研究，我们提出了TMS诱发EEG的全脑兴奋性指标，脑耦合和脑网络指标，该指标与脑状态相关。我们将这些脑状态评估方法运用到了意识障碍患者的临床诊断和治疗评估中。结合TMS诱发EEG的脑状态评估，我们成功地将微意识状态的患者从植物状态的患者中挑选出来，为后期的临床医疗提供技术支撑。同时，我们用该方法评估意识障碍患者在重复经颅磁刺激中的反应，监测到部分患者大脑在重复经颅磁刺激治疗中的变化，为将重复经颅磁刺激技术引入到意识障碍康复领域提供了客观的证据。在本项目支撑下，我们已经获得了50例意识障碍患者的TMS诱发EEG信号，为我们下一步提出临床用意识障碍患者脑状态检测系统提供了支撑。