Decoding imagined speech with ear-EEG based brain-computer interface

Abstract

This dissertation investigates how effective electroencephalography (EEG) signals collected from aroundthe user’s ears (ear-EEG) when used for a speech-imagery-based brain-computer interface (BCI) system.A low-cost wearable ear-EEG acquisition device was developed and used in comparison with a conventional 32-channel scalp-EEG collection counterpart in a multi-class speech imagery classification taskusing several machine learning models. Data was collected from ten subjects in an experiment consistingof six sessions spanning three days. The experiment involved imagining four speech commands (’Left,’’Right,’ ’Forward,’ and ’Go back’) and staying in a rest condition. The classification accuracy of oursystem is significantly above the chance level (20%) for both ear-EEG and scalp-EEG. The best performing classification model result averaged across all ten subjects is 38.2% and 43.1% with a maximum(max) of 43.8% and 55.0% for ear-EEG and scalp-EEG, respectively. Seven out of ten subjects showno significant difference between the speech-imagery classification performance when using ear-EEG andscalp-EEG. The results indicate that ear-EEG has great potential as an alternative to the scalp-EEGacquisition method for speech-imagery monitoring. We believe that the merits and feasibility of bothspeech imagery and ear-EEG acquisition in the proposed system will accelerate the development of theBCI system for daily-life use.