Evoked response potential training on a consumer EEG headset

Abstract

Neurofeedback paradigms are of increasing interest, both for brain-computer interfaces and for training purposes. In the current study we examine multi-session training in a P3 speller application which uses P300 evoked response potentials to identify the letter of a word a participant is attempting to spell. Because neurofeedback training is most effective over time, there is an advantage to using low-cost hardware that can be used in the home rather than expensive research-grade equipment. This study is performed on an Emotiv Epoc EEG headset and some of the problems and workarounds inherent in this are discussed. This headset is sensitive enough to detect evoked response potentials and evidence of a training effect is found. Motivational aspects of ERP training are discussed and an alternate P300 training paradigm is presented.