Deep learning for brain computer interfaces

Abstract

From playing games with just the mind to capturing and re-constructing dreams, Brain computer Interfaces (BCIs) have turned fiction into reality. It has set new standards in the world of prosthetics, be it hearing aids or prosthetic arms, legs or vision, helping paralyzed or completely locked-in users. Not only can one get a visual imprint of their own brain activity but the future of BCI will make sharing someone else’s experience possible. The entire functioning of the BCI can be segmented into acquiring the signals, processing it, translation of signals, device that gives the output and the protocol in operation. The translation algorithms can be classical statistical analysis or non-linear methods such as neural networks. Deep learning might serve as one of the translation algorithms that converts the raw signals from the brain into commands that the output devices follow. This chapter aims to give an insight into the various deep learning algorithms that have served in BCI’s today and helped enhance their performances.