Convolutional Neural Networks for a Cursor Control Brain Computer Interface

International audienceA Brain-Computer Interface (BCI) platform can be utilized by a patient to control an external device without making any overt movements. This can be beneficial to a variety of patients who suffer from paralysis, loss of limb, or neurodegenerative diseases. We decode brain signals using EEG during imagined body kinematics to control an on-screen cursor. Convolutional neural networks (CNNs) are already a popular choice for image-based learning problems and are useful in EEG applications. The major advantage of CNNs is that they can generate features from the signal automatically and do not require as much domain driven feature engineering as a traditional machine learning approach. We implement a CNN to perform multivariate regression over the EEG signal to predict intended cursor velocity