CoCo Brain Channel
With the recent advances in EEG technology and the popularization of low-cost mobile EEG devices, brain-computer interface systems and neurofeedback tools become more accessible. Moreover, machine learning algorithms expand the scope of problems that can be tackled with these techniques, including classification of mental states. Real-time EEG signal processing is increasingly popular in the context of digital arts projects powered by a neuroaesthetic approach. CoCo Brain Channel is one such project : designed to use real-time processing of EEG signal in order to generate a musical environment, it provides the user with a mean of hearing and controlling his own brain activity. Thus, we hook-up a commercial mobile EEG device to a music generation algorithm built in PureData. Features are computed in real-time from the EEG signal and sent both to PureData for sonification, and to a Python (Scikit-learn) script for custom classification. The script uses a boosted random forest classification algorithm to compute predictions that are sent to PureData. The result is a continuous musical soundscape based on modulations of the EEG signals. Conceptually, EEG-based real-time sonification allows the user to observe his mental-state throughout this creative process. The CoCo Brain Channel set-up is a tool that allows exploration of the neural correlates of various mental states in a playful and artistic manner. Improvements and possible applications for basic research will be discussed.