Auditory Imagery Classification with a Non-Invasive Brain Computer Interface

Abstract

BrainComputer Interfaces using motor imagery have been exploredduringmany years. Auditory imagery,nonetheless, has been a rarely explored approach, buta one that might open new possibilities in musical interface, communication and speech synthesis.Using a non-invasive BCI,open source software, andau-ditory imagery ofwhite noise, we testedseveral classification algorithms todetermine the accuracy and usefulnessof using non-motor imagery EEG signals.We tested 15 healthy adults with a 6 electrode EEG setup and the open source platform OpenVibe. Our results showthat using a Support Vector Machine classifier and our experimental setup, we could achieve up to93%accuracy in white noiseimagery in our subjects. Lin-ear Discriminant Analysis and MultiLayer Perceptron setups yielded accuracy of 73-76%.