A REVIEW OF RESEARCHES ON MACHINE LEARNING OF STEADY STATE VISUAL EVOKED POTENTIAL

Steady state visual evoked potential (SSVEP) has become one of the major paradigms in BCI research due to its high signal-to-noise ratio and high information transfer rate. Using algorithm to recognize and extract the features of SSVEP signal is the key of SSVEP system research. However, the current researches lack of SSVEP algorithm review. For this problem, this paper focused on the progress of SSVEP machine learning in recent years. From the perspective of machine learning, algorithms were divided into supervised learning and unsupervised learning. This paper explained the fundamental such as canonical correlation analysis and convolutional neural networks. This paper summarized the shortcomings of current SSVEP algorithm in practical application and discussed the opportunities and challenges faced by SSVEP.