The Survey of Image Generation from EEG Signals based on Deep Learning

Abstract

China has become a high-risk region of stroke. Most patients with stroke suffer regular bouts of post-stroke limb dyskinesia. Nowadays, there isn't an effective treatment for these patients. Brain computer interface (BCI) establishes a new pathway to connect human brains and device, which provide an innovation method to repair the human brain nervous systems through rehabilitation training. However, one of the mainly brain activity recordings, Electroencephalogram (EEG), cannot be represented accurately by other algorithms. With the development of deep learning techniques, the topic of EEG signals' representation by image generation technique has become an important research area. This paper we introduced the basic concepts of BCI systems first, then we give a survey of image generation techniques from EEG signals. At last, we proposed an experimental scheme of dataset establishment which is used for post-stroke patients with upper limb dyskinesia