Real-time Naive Learning of Neural Correlates in ECoGElectrophysiology

Abstract

Brain Computer Interfaces (BCI) seek to measure brain signals in order to control computational or robotic devices, with important applications to motor disability. Electrocorticography (ECoG) is an emerging signal platform for long term implantation of a brain signal recording device, but current approaches rely heavily on screening tasks and trained technicians to find and specify repeatable features in the ECoG signal. Here we explore unsupervised approaches to reducing the ECoG signal stream into a few components that correspond most directly to neural patterns that correlate to subject task performance (neural correlates). We report on the development of a real-time feedback system we call the "Brain Mirror" which is based on the real time, incremental learning of a Deep Belief Network. On real patient data, we demonstrate that the components learned online with Deep Belief Networks have higher correlations with neural patterns than PCA.