BlinkFruity: A Real-Time EEG Based Neurofeedback Game for Brain-Computer Interface

Abstract

The Brain-Computer Interface (BCI) is a communication channel between the brain and the computer. It works by detecting the neural signal of brain activity. Brain signals can be detected with various BCI methods that include observing the changes in magnetic fields due to electric currents, changes associate with blood flow and neural electrical activity. Based on the selection of the required area of brain signal and the BCI method, the signal acquisition process might require from non-invasive method to an invasive (surgical) method. Despite invasive methods require a surgical procedure to implement brain signal recording implants, brain signals recorded via invasive methods offers unparalleled spatial resolution with a very low signal to noise ratio (SNR). On the contrary, non-invasive methods are easy to implement and provides good temporal resolution with high SNR. Among many non-invasive brain signal recording methods, Electroencephalography (EEG) is one of the most popular methods. EEG controlled applications widely range from strictly medical to non-medical applications. Non-medical applications can not only be used for entertainment purposes but also can help a subject to experience BCI application, achieve better control with rehabilitation systems and can be a strong motivation to practise the BCI system. Brain signals recorded via EEG are weak and contain several artefacts like muscle movements, cardiac, eye blink, power source, and amplitude artefacts. Although eye blink is considered as one of the strongest artefacts, but it can be used to drive BCI enabled applications. With this in mind, this study uses eye blink as a control signal to play the BlinkFruity game, where users collect fruits into a basket using eye blink only. To attain this objective, at first, the brain activity was recorded using OpenBCI device and 500ms window data was taken to process in real-time. Then the notch filter was applied to remove powerline noise. The eye blink was detected by using the signal thresholding method reading from EEG data. Blink detection average accuracy of 84.8% was obtained using blink control applying on subjects. The primary objective of this study is to design a simple BCI enabled system for users who are experiencing BCI for the first time and find it interesting. Then evaluate the proposed system and user’s experience. Even though in this experiment, blink has been used for experiment purpose, there are several areas where blink can be used as home automation, rehabilitation and augmented mobile application experience. Moreover, findings from this study can be resourceful and enhance our understanding and capacity for developing BCI application.