EEG data for motor imagery brain-computer interface using low-cost equipment

Abstract

EEG-based brain-computer interfaces (BCI) for motor imagery recognition can be used in many applications, including prosthesis control, post-stroke motor rehabilitation, communication, and videogames. Such BCIs usually need to be calibrated with EEG data before being used. The calibration can use data from either a single person, the same person who will use the equipment, or a group of different people. However, although BCIs are increasingly used in research and real-world problems, high equipment costs prevent their popularization in personal use applications. For this reason, there are many ongoing efforts to create more affordable BCI devices. Nevertheless, most public datasets for motor imagery EEG-BCIs still use expensive equipment. Therefore, our work presents a dataset for EEG-based motor imagery BCIs focused on personal use applications. Using a low-cost 16-electrode EEG OpenBCI Cyton+Daisy Biosensing Board, we recorded the brain signals of 6 subjects while they imagined the movements of their hands, resulting in a dataset containing 960 trials of left and right-hand motor imagery. This dataset can be used to calibrate BCIs using similar low-cost equipment as well as study the signals generated by such equipment.