An EEG dataset for studying asynchronous steady-state visual evoked potential (SSVEP) based brain computer interfaces

Compared with the commonly used synchronous brain-computer interface (BCI), the asynchronous BCI is a more flexible and natural way to control the real-world robotic devices. The major difficulty of building a robust asynchronous BCI lies in the discrimination between control states (CSs) and non-control states (NSs). This article presents an open-source 63-channel electroencephalogram (EEG) dataset of 24 subjects for asynchronous steady-state visual evoked potential (SSVEP)-BCI research. The data was recorded from an SSVEP based CS task and three different types of NS tasks, namely NS1, NS2 and NS3. The dataset was evaluated using three processes, including analysis of temporal waveform, amplitude spectrum and signal-to-noise (SNR), recognition of SSVEP frequencies, and classification between CS and NS. The dataset can be used to support studies of asynchronous classification, NS detection, and discrimination between different NSs.