Processing EEG signals towards the construction of a user experience assessment method

Abstract

This paper proposes a neural network to identify pleasant and unpleasant emotions from recorded electroencephalography (EEG) signals, towards the construction of a method to assess user experience (UX). EEG signals were obtained with an Emotiv EEG device. The input data was recorded during the presentation of visual stimulus that induce emotions known a priori. The EEG signals recorded were preprocessed to enhance the differences and then used to train and validate a Patternet neural network. The results indicate that the neural network provides an accurate rate of 99.61% for 258 preprocessed signals.