如何识别人的情绪是情感脑机交互研究的一个重要课题。然而，目前情绪识别模型的准确率还比较低，还无法满足情感脑机交互实际应用的要求，人们在构建情绪识别模型时尚未充分考虑人自身及其文化差异对情绪识别的影响。本项目利用脑电和眼动信号探索情绪识别的年龄差异、性别差异和文化差异，拟开展的研究工作包括以下三个方面：1）通过五类情绪识别任务，探索脑电信号关键频段、关键脑区、关键脑网络和眼动信号关键特征的年龄差异、性别差异和文化差异；2）研究可有效减少脑电和眼动信号年龄差异、性别差异和文化差异的域适应和域泛化学习算法，使跨被试的五类情绪识别模型的准确率有大幅度的显著提升；3）针对脑电信号采集成本高和样本量少的问题，研究新的生成式对抗网络模型，生成有效的脑电和眼动数据，以增加训练样本，提升情绪识别模型的性能。本项目对研究大脑情绪处理的机制和探索情感脑机交互技术在医疗、交通和国防等领域的实际应用都具有重要意义。

How to recognize people’s emotion has always been a key issue in affective brain-computer interaction research. However, the accuracy of emotion recognition models is still low currently, and it can not meet the requirements of affective brain-computer interaction application. People have not fully considered the influence of human and their culture differences on emotion recognition when constructing emotion recognition models. Therefore, this project aims to utilize EEG and eye movement signals to explore age, gender and cultural differences in emotion recognition. The research work to be carried out includes the following three aspects: 1) Explore age, gender and cultural differences in EEG signal’s key frequency bands, key brain regions, key brain networks and eye movement signal’s key features through five emotion classification task; 2) Research domain adaptation and domain generalization learning algorithm that can effectively reduce age, gender and cultural differences in EEG and eye movement signals, so that the classification accuracy of cross-subject five emotion classification model can be significantly improved ; 3) For the problem of high cost in EEG signal acquisition and insufficient samples, research new generative adversarial network model, generate effective EEG and eye movement data to increase the number of training samples, and improve the performance of emotion recognition model. This study is of great significance for studying the mechanism of the brain’s emotion processing and exploring application of affective brain-computer interaction in the fields of medical care, transportation and national defense.