基于脑电信号分析的不确定决策神经机制研究

Uncertainty decision-making is an important research direction of neuroeconomics. Studies on how the brain processes uncertainty and how to make decisions under uncertain conditions, is of great significance for understanding the neural mechanisms related to decision-making. However, at the current stage of research on uncertain decision-making based on EEG, the research on ERP components and their differences in the early stage of decision-making, the difference of EEG rhythm components during uncertain decision-making processing, and the dynamic changes of EEG-based brain networks are still insufficient. This project will study the characteristics of uncertain decision-making based on EEG signals, and the main research contents include: (1) exploring early ERP components and their differences during uncertainty decision-making; (2) investigating the EEG rhythm components related to uncertainty decision-making and its dynamic characteristics；(3) analyzing of the dynamic changes of the brain function network during the decision-making processes. The research of this project is expected to further reveal the neural mechanism of decision-making, and to understand and explain the individual's decision-making behavior on the scientific level, which has important theoretical and practical significances.

不确定决策作为神经经济学领域的重要研究方向之一，研究大脑如何加工不确定性以及如何不确定情境下做出决策，对于理解决策相关的神经机理具有重要意义。然而，现阶段基于脑电的不确定决策研究中，决策早期ERP成分及其差异、脑电节律信号在不确定决策过程中的差异、基于EEG的脑网络动态变化等方面的研究尚不充分。本项目将基于脑电信号对不确定决策的特性进行研究，主要研究内容包括：（1）探索不确定决策过程中的早期ERP成分及其差异；（2）研究不确定决策过程相关的脑电节律成分及其动态变化过程；（3）分析决策过程中脑功能网络的动态变化特性。本项目的研究有望进一步揭示不确定决策背后的神经机制，并从科学层面上理解和解释人类个体的决策行为，具有重要的理论意义和实际意义。

为了进一步更有效的提取脑电信号的特征，提升脑电信号的分类和解码性能，以进一步提升脑信号分析在神经决策、疾病诊断方面的应用，本项目在以下几个方面进行研究：1）基于时空滤波的导联选择算法用于单试次EEG分类；2）基于正交小波分解的算法并用于运动想象脑电信号分类；3）基于数据合成的卷积编码器-解码器网络进行电磁源成像；4）提出了一种基于黎曼几何的运动想象分类方法；5） 开发了通过双水平多模态刺激和稀疏空间模式解码方法以增强下肢运动想象能力的系统；6）提出一种基于逻辑回归的算法以实现时空滤波器和分类器联合优化。我们提出的相关算法有助于提升脑电信号的分类和解码的性能，其中一部分研究工作已经应用到系统设计研究中。此部分算法研究工作为后续进行脑电的相关领域研究奠定了坚实的理论基础。