Using general-purpose graphic processing units for BCI systems

BioMEMS electrode array fabrication techniques are used to develop high-density arrays with hundreds of channels. However, it was previously impossible to process more than a fraction of these channels real-time for online BCI experiments due to computational resource restraints. It is now possible to use graphics processing units (GPUs), which can have several hundred processing cores each, to processes large amounts of data quickly. This paper summarizes advances in using GPUs for BCI processing for EEG, ECoG, and micro-electrode systems, with speedups of more than 30 times that of current state-of-the-art CPU-based BCI implementations.