这篇文章是一篇关于用于EEG分类的深度学习方法的综述:发表在Journal of Neural Engineering
Which EEG classification tasks have been explored with deep learning?
共有六种分类的任务:运动想象任务(Motor imagery tasks)、脑力工作任务(Mental workload tasks)、癫痫检测任务(Seizure detection tasks)、睡眠阶段评分任务(Sleep stage scoring tasks)、事件相关电位任务(Event related potential tasks)
Preprocessing methods
关于伪影去除的策略:不同的分类任务有不同的滤波范围,伪影去除的方法有四种:手动去除(独立主成分分析ICA或离散小波变换DWT的方法)、自动去除、不处理
What input formulations have been used for training the deep networks?
calculated features (计算特征), images (图像), the signal values (信号值)