作为人类高级认知加工的中枢，大脑以不到3%的体积占比却消耗了人体大约20%左右的能量。这些能量如何供应大脑完成各种神经活动，进而实现各种复杂的认知功能，一直是研究者们重点探索的方向之一。发生在大脑局部的神经血管耦合现象可以为研究者了解大脑的高级认知过程提供帮助。为了探究生物体内的神经血管耦合现象，需要对神经元活动信号与血流动力学信号进行同步采集。针对神经元活动信号，目前应用最广泛的测量技术是脑电（Electroencephalogram，EEG）技术，而对于血流动力学信号，一般则选用功能近红外光谱（Functional Near-Infrared Spectroscopy，fNIRS）或功能磁共振技术（Functional Magnetic Resonance Imaging，fMRI）进行测量。EEG-fNIRS联合采集系统因为其在成本、测量原理、便携性、被试服从度的优良特性得到了越来越广泛的应用。但在得到EEG-fNIRS联合采集数据后，如何将两者的数据特征进行融合，进而对认知过程中的神经血管耦合现象进行探究成为了新的问题。本论文发展了一种新的EEG-fNIRS数据融合算法，将EEG信号所代表的神经活动与fNIRS代表的血流动力学活动在数据层面联结起来，从数据角度为研究者认识高级认知过程中的神经血管耦合机制提供了新的思路。在发展了一种新的联合分析方法后，本文设计了两个经典的认知实验，探究任务负荷增加时神经血管耦合机制特征的变化规律：研究一基于数字刺激的nback实验对被试在不同任务难度下的工作记忆过程进行探索，在同步采集了三十名被试EEG信号与fNIRS信号后，对被试在不同任务难度下的EEG信号特征、fNIRS信号特征、EEG-fNIRS联合分析特征进行了阐述，联合分析的结果进一步验证了，随着任务难度的上升，自顶向下的控制机制以前额叶区域alpha节律的抑制驱动了血氧信号的进一步激活，以更好地适应任务难度的变化。研究二通过视觉搜索范式刺激方向是否随机以区分不同的任务难度，探索了不同难度下被试在空间注意分配过程的神经机制差异。在实验过程中采集了五十名被试的EEG信号及fNIRS信号，探索了被试在注意力分配期间的神经机制及血氧信号变化的规律，随着任务难度的上升，对于注意力分配的要求也随之提高，而在这一过程中，自顶向下的调节机制以顶叶及背外侧前额叶区域beta节律的增强驱动了血氧信号的进一步激活，以此适应了任务难度变化下的注意力需求。综上所述，本文发展了一种新的EEG-fNIRS联合分析方法，可以直观有效的对两种信号间的联合变化规律进行度量；基于EEG-fNIRS联合方法对任务负荷变化下的神经血管耦合机制进行了探索，为后续EEG-fNIRS联合采集、分析技术的进一步发展打下了基础，也进一步探索任务负荷导致的认知机制变化。

The brain, as the center of human advanced cognitive processing, accounts for less than 3% of the body's volume but consumes about 20% of its energy. Researchers have been focusing on exploring how the brain is supplied with energy to carry out various neural activities and achieve complex cognitive functions. The neurovascular coupling phenomenon that occurs in localized areas of the brain can help researchers understand the advanced cognitive processes of the brain. To investigate the neurovascular coupling phenomenon in biological organisms, it is necessary to synchronously acquire neuronal activity signals and hemodynamic signals. The most widely used measurement technique for neuronal activity signals is electroencephalography (EEG), while functional near-infrared spectroscopy (fNIRS) or functional magnetic resonance imaging (fMRI) is generally used for hemodynamic signals. The EEG-fNIRS joint acquisition system has been increasingly used due to its excellent characteristics in cost, measurement principle, portability, and subject compliance. However, after obtaining the EEG-fNIRS joint acquisition data, how to fuse the data features of both and explore the neurovascular coupling phenomenon in cognitive processes has become a new problem. First, this paper develops a new EEG-fNIRS data fusion algorithm that links the neural activity represented by the EEG signal and the hemodynamic activity represented by the fNIRS signal at the data level, providing a new idea for researchers to understand the neurovascular coupling mechanism in advanced cognitive processes from a data perspective. Subsequently, after developing a new joint analysis method, this paper designs two classic cognitive experiments to explore the changes in neurovascular coupling mechanism characteristics as the task load increases: In the first experiment, a digit stimulus-based n-back experiment was conducted to explore the working memory process of subjects under different task difficulties. Thirty subjects were synchronously measured for EEG signals and fNIRS signals, and the characteristics of EEG signals, fNIRS signals, and EEG-fNIRS joint analysis features under different task difficulties were described. The joint analysis results further verified that as the task difficulty increased, the top-down control mechanism suppressed alpha rhythms in the frontal lobe area to drive further activation of the blood oxygen signal, thereby better adapting to changes in task difficulty. In the second experiment, a visual search paradigm was used to differentiate task difficulties based on whether the direction of the stimulus was random, exploring the neural mechanisms differences in spatial attention allocation of subjects under different difficulties. EEG signals and fNIRS signals were collected from fifty subjects during the experiment to explore the neural mechanisms and blood oxygen signal changes during attention allocation. As the task difficulty increased, the requirements for attention allocation also increased, and in this process, the top-down regulation mechanism in the parietal lobe and the dorsolateral prefrontal cortex drove the further activation of the blood oxygen signal through beta rhythms, adapting to the changes in attentional demands under task difficulty. In summary, this study has developed a novel EEG-fNIRS joint analysis method that allows for intuitive and effective measurement of the combined variations between the two signals. Based on the EEG-fNIRS joint approach, the neurovascular coupling mechanism under task load variations was explored, laying the foundation for further development of EEG-fNIRS joint data collection and analysis techniques. Additionally, it further investigates cognitive mechanism changes induced by task load.